/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1680 by zherczeg, Tue Feb 14 08:48:18 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.41
8    ------------
9    
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
12    
13    2. Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14    mode with --only-matching matched several lines, it restarted scanning at the
15    next line instead of moving on to the end of the matched string, which can be
16    several lines after the start.
17    
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
19    
20    
21    Version 8.40 11-January-2017
22    ----------------------------
23    
24    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
25        the match extended over a line boundary.
26    
27    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
28        MSVC static compilation, putting the first patch under a new option.
29    
30    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
31    
32    4.  Ignore "show all captures" (/=) for DFA matching.
33    
34    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
35    
36    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
37        without PCRE_UCP set, a negative character type such as \D in a positive
38        class should cause all characters greater than 255 to match, whatever else
39        is in the class. There was a bug that caused this not to happen if a
40        Unicode property item was added to such a class, for example [\D\P{Nd}] or
41        [\W\pL].
42    
43    7.  When pcretest was outputing information from a callout, the caret indicator
44        for the current position in the subject line was incorrect if it was after
45        an escape sequence for a character whose code point was greater than
46        \x{ff}.
47    
48    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
49        the conditional was interpreted as a reference to capturing group 1 instead
50        of a test for recursion. Any group whose name began with R was
51        misinterpreted in this way. (The reference interpretation should only
52        happen if the group's name is precisely "R".)
53    
54    9.  A number of bugs have been mended relating to match start-up optimizations
55        when the first thing in a pattern is a positive lookahead. These all
56        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
57    
58        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
59            both an initial 'X' and a following 'X'.
60        (b) Some patterns starting with an assertion that started with .* were
61            incorrectly optimized as having to match at the start of the subject or
62            after a newline. There are cases where this is not true, for example,
63            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
64            start with spaces. Starting .* in an assertion is no longer taken as an
65            indication of matching at the start (or after a newline).
66    
67    
68    Version 8.39 14-June-2016
69    -------------------------
70    
71    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
72        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
73        misbehaved. This bug was found by the LLVM fuzzer.
74    
75    2.  Similar to the above, if an isolated \E was present between an item and its
76        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
77        bug was found by the LLVM fuzzer.
78    
79    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
80        working correctly in UCP mode.
81    
82    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
83        was set when the pmatch argument was NULL. It now returns REG_INVARG.
84    
85    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
86    
87    6.  An empty \Q\E sequence between an item and its qualifier caused
88        pcre_compile() to misbehave when auto callouts were enabled. This bug was
89        found by the LLVM fuzzer.
90    
91    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
92        space or a #-type comment that was followed by (?-x), which turns off
93        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
94        pcre_compile() assumed that (?-x) applied to the whole pattern and
95        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
96    
97    8.  A call of pcre_copy_named_substring() for a named substring whose number
98        was greater than the space in the ovector could cause a crash.
99    
100    9.  Yet another buffer overflow bug involved duplicate named groups with a
101        group that reset capture numbers (compare 8.38/7 below). Once again, I have
102        just allowed for more memory, even if not needed. (A proper fix is
103        implemented in PCRE2, but it involves a lot of refactoring.)
104    
105    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
106        start of the match to be earlier than the end.
107    
108    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
109    
110    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
111        assertion, caused pcretest to generate incorrect output, and also to read
112        uninitialized memory (detected by ASAN or valgrind).
113    
114    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
115        nested set of parentheses of sufficient size caused an overflow of the
116        compiling workspace (which was diagnosed, but of course is not desirable).
117    
118    14. And yet another buffer overflow bug involving duplicate named groups, this
119        time nested, with a nested back reference. Yet again, I have just allowed
120        for more memory, because anything more needs all the refactoring that has
121        been done for PCRE2. An example pattern that provoked this bug is:
122        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
123        registered as CVE-2016-1283.
124    
125    15. pcretest went into a loop if global matching was requested with an ovector
126        size less than 2. It now gives an error message. This bug was found by
127        afl-fuzz.
128    
129    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
130        ("assertion expected") when (?(?C) was not followed by an opening
131        parenthesis.
132    
133    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
134        actually affect anything, by sheer luck.
135    
136    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
137        static compilation.
138    
139    19. Modified the RunTest script to incorporate a valgrind suppressions file so
140        that certain errors, provoked by the SSE2 instruction set when JIT is used,
141        are ignored.
142    
143    20. A racing condition is fixed in JIT reported by Mozilla.
144    
145    21. Minor code refactor to avoid "array subscript is below array bounds"
146        compiler warning.
147    
148    22. Minor code refactor to avoid "left shift of negative number" warning.
149    
150    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
151        UCP support.
152    
153    24. Refactor to avoid compiler warnings in pcrecpp.cc.
154    
155    25. Refactor to fix a typo in pcre_jit_test.c
156    
157    26. Patch to support compiling pcrecpp.cc with Intel compiler.
158    
159    
160    Version 8.38 23-November-2015
161    -----------------------------
162    
163    1.  If a group that contained a recursive back reference also contained a
164        forward reference subroutine call followed by a non-forward-reference
165        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
166        compile correct code, leading to undefined behaviour or an internally
167        detected error. This bug was discovered by the LLVM fuzzer.
168    
169    2.  Quantification of certain items (e.g. atomic back references) could cause
170        incorrect code to be compiled when recursive forward references were
171        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
172        This bug was discovered by the LLVM fuzzer.
173    
174    3.  A repeated conditional group whose condition was a reference by name caused
175        a buffer overflow if there was more than one group with the given name.
176        This bug was discovered by the LLVM fuzzer.
177    
178    4.  A recursive back reference by name within a group that had the same name as
179        another group caused a buffer overflow. For example:
180        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
181    
182    5.  A forward reference by name to a group whose number is the same as the
183        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
184        a buffer overflow at compile time. This bug was discovered by the LLVM
185        fuzzer.
186    
187    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
188        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
189    
190    7.  Another buffer overflow bug involved duplicate named groups with a
191        reference between their definition, with a group that reset capture
192        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
193        fixed by always allowing for more memory, even if not needed. (A proper fix
194        is implemented in PCRE2, but it involves more refactoring.)
195    
196    8.  There was no check for integer overflow in subroutine calls such as (?123).
197    
198    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
199        being treated as a literal 'l' instead of causing an error.
200    
201    10. There was a buffer overflow if pcre_exec() was called with an ovector of
202        size 1. This bug was found by american fuzzy lop.
203    
204    11. If a non-capturing group containing a conditional group that could match
205        an empty string was repeated, it was not identified as matching an empty
206        string itself. For example: /^(?:(?(1)x|)+)+$()/.
207    
208    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
209        \a and \e in test subject lines.
210    
211    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
212        instead of the EBCDIC value.
213    
214    14. The handling of \c in an EBCDIC environment has been revised so that it is
215        now compatible with the specification in Perl's perlebcdic page.
216    
217    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
218        ASCII/Unicode. This has now been added to the list of characters that are
219        recognized as white space in EBCDIC.
220    
221    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
222        error (correctly) when used outside a class, but did not give an error
223        within a class.
224    
225    17. \h within a class was incorrectly compiled in EBCDIC environments.
226    
227    18. A pattern with an unmatched closing parenthesis that contained a backward
228        assertion which itself contained a forward reference caused buffer
229        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
230    
231    19. JIT should return with error when the compiled pattern requires more stack
232        space than the maximum.
233    
234    20. A possessively repeated conditional group that could match an empty string,
235        for example, /(?(R))*+/, was incorrectly compiled.
236    
237    21. Fix infinite recursion in the JIT compiler when certain patterns such as
238        /(?:|a|){100}x/ are analysed.
239    
240    22. Some patterns with character classes involving [: and \\ were incorrectly
241        compiled and could cause reading from uninitialized memory or an incorrect
242        error diagnosis.
243    
244    23. Pathological patterns containing many nested occurrences of [: caused
245        pcre_compile() to run for a very long time.
246    
247    24. A conditional group with only one branch has an implicit empty alternative
248        branch and must therefore be treated as potentially matching an empty
249        string.
250    
251    25. If (?R was followed by - or + incorrect behaviour happened instead of a
252        diagnostic.
253    
254    26. Arrange to give up on finding the minimum matching length for overly
255        complex patterns.
256    
257    27. Similar to (4) above: in a pattern with duplicated named groups and an
258        occurrence of (?| it is possible for an apparently non-recursive back
259        reference to become recursive if a later named group with the relevant
260        number is encountered. This could lead to a buffer overflow. Wen Guanxing
261        from Venustech ADLAB discovered this bug.
262    
263    28. If pcregrep was given the -q option with -c or -l, or when handling a
264        binary file, it incorrectly wrote output to stdout.
265    
266    29. The JIT compiler did not restore the control verb head in case of *THEN
267        control verbs. This issue was found by Karl Skomski with a custom LLVM
268        fuzzer.
269    
270    30. Error messages for syntax errors following \g and \k were giving inaccurate
271        offsets in the pattern.
272    
273    31. Added a check for integer overflow in conditions (?(<digits>) and
274        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
275        fuzzer.
276    
277    32. Handling recursive references such as (?2) when the reference is to a group
278        later in the pattern uses code that is very hacked about and error-prone.
279        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
280        give an internal error if it is obvious that compiling has gone wrong.
281    
282    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
283        This issue was found by Karl Skomski with a custom LLVM fuzzer.
284    
285    34. The JIT compiler should restore the control chain for empty possessive
286        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
287    
288    35. Match limit check added to JIT recursion. This issue was found by Karl
289        Skomski with a custom LLVM fuzzer.
290    
291    36. Yet another case similar to 27 above has been circumvented by an
292        unconditional allocation of extra memory. This issue is fixed "properly" in
293        PCRE2 by refactoring the way references are handled. Wen Guanxing
294        from Venustech ADLAB discovered this bug.
295    
296    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
297        with a custom LLVM fuzzer.
298    
299    38. Fixed a corner case of range optimization in JIT.
300    
301    39. An incorrect error "overran compiling workspace" was given if there were
302        exactly enough group forward references such that the last one extended
303        into the workspace safety margin. The next one would have expanded the
304        workspace. The test for overflow was not including the safety margin.
305    
306    40. A match limit issue is fixed in JIT which was found by Karl Skomski
307        with a custom LLVM fuzzer.
308    
309    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
310        work under Windows. (Why has it taken so long for anyone to notice?)
311    
312    42. In a character class such as [\W\p{Any}] where both a negative-type escape
313        ("not a word character") and a property escape were present, the property
314        escape was being ignored.
315    
316    43. Fix crash caused by very long (*MARK) or (*THEN) names.
317    
318    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
319        by a single ASCII character in a class item, was incorrectly compiled in
320        UCP mode. The POSIX class got lost, but only if the single character
321        followed it.
322    
323    45. [:punct:] in UCP mode was matching some characters in the range 128-255
324        that should not have been matched.
325    
326    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
327        class, all characters with code points greater than 255 are in the class.
328        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
329        such as \w are turned into Unicode properties), wide characters were not
330        correctly handled, and could fail to match.
331    
332    
333    Version 8.37 28-April-2015
334    --------------------------
335    
336    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
337        for those parentheses to be closed with whatever has been captured so far.
338        However, it was failing to mark any other groups between the hightest
339        capture so far and the currrent group as "unset". Thus, the ovector for
340        those groups contained whatever was previously there. An example is the
341        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
342    
343    2.  If an assertion condition was quantified with a minimum of zero (an odd
344        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
345    
346    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
347        unrecognized modifier, a crash could occur.
348    
349    4.  An attempt to do global matching in pcretest with a zero-length ovector
350        caused a crash.
351    
352    5.  Fixed a memory leak during matching that could occur for a subpattern
353        subroutine call (recursive or otherwise) if the number of captured groups
354        that had to be saved was greater than ten.
355    
356    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
357        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
358        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
359    
360    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
361        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
362    
363    8. When a pattern is compiled, it remembers the highest back reference so that
364       when matching, if the ovector is too small, extra memory can be obtained to
365       use instead. A conditional subpattern whose condition is a check on a
366       capture having happened, such as, for example in the pattern
367       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
368       setting the highest backreference number. This mattered only if pcre_exec()
369       was called with an ovector that was too small to hold the capture, and there
370       was no other kind of back reference (a situation which is probably quite
371       rare). The effect of the bug was that the condition was always treated as
372       FALSE when the capture could not be consulted, leading to a incorrect
373       behaviour by pcre_exec(). This bug has been fixed.
374    
375    9. A reference to a duplicated named group (either a back reference or a test
376       for being set in a conditional) that occurred in a part of the pattern where
377       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
378       to be incorrectly calculated, leading to overwriting.
379    
380    10. A mutually recursive set of back references such as (\2)(\1) caused a
381        segfault at study time (while trying to find the minimum matching length).
382        The infinite loop is now broken (with the minimum length unset, that is,
383        zero).
384    
385    11. If an assertion that was used as a condition was quantified with a minimum
386        of zero, matching went wrong. In particular, if the whole group had
387        unlimited repetition and could match an empty string, a segfault was
388        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
389        assertions to be quantified, but not if they are being used as conditions,
390        so the above pattern is faulted by Perl. PCRE has now been changed so that
391        it also rejects such patterns.
392    
393    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
394        failed to allow the zero-repeat case if pcre2_exec() was called with an
395        ovector too small to capture the group.
396    
397    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
398        Red Hat Product Security:
399    
400        (a) A crash if /K and /F were both set with the option to save the compiled
401        pattern.
402    
403        (b) Another crash if the option to print captured substrings in a callout
404        was combined with setting a null ovector, for example \O\C+ as a subject
405        string.
406    
407    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
408        forward reference repeated a large (but limited) number of times within a
409        repeated outer group that has a zero minimum quantifier, caused incorrect
410        code to be compiled, leading to the error "internal error:
411        previously-checked referenced subpattern not found" when an incorrect
412        memory address was read. This bug was reported as "heap overflow",
413        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
414        CVE-2015-2325.
415    
416    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
417        call within a group that also contained a recursive back reference caused
418        incorrect code to be compiled. This bug was reported as "heap overflow",
419        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
420        number CVE-2015-2326.
421    
422    24. Computing the size of the JIT read-only data in advance has been a source
423        of various issues, and new ones are still appear unfortunately. To fix
424        existing and future issues, size computation is eliminated from the code,
425        and replaced by on-demand memory allocation.
426    
427    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
428        adjacent to the end of the range, and the range contained characters with
429        more than one other case, caused incorrect behaviour when compiled in UTF
430        mode. In that example, the range a-j was left out of the class.
431    
432    26. Fix JIT compilation of conditional blocks, which assertion
433        is converted to (*FAIL). E.g: /(?(?!))/.
434    
435    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
436        discovered by the LLVM fuzzer.
437    
438    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
439        when this assertion was used as a condition, for example (?(?!)a|b). In
440        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
441        error about an unsupported item.
442    
443    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
444        possessification code could take exponential time to complete. A recursion
445        depth limit of 1000 has been imposed to limit the resources used by this
446        optimization.
447    
448    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
449        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
450        because \S ensures they are all in the class. The code for doing this was
451        interacting badly with the code for computing the amount of space needed to
452        compile the pattern, leading to a buffer overflow. This bug was discovered
453        by the LLVM fuzzer.
454    
455    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
456        other kinds of group caused stack overflow at compile time. This bug was
457        discovered by the LLVM fuzzer.
458    
459    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
460        between a subroutine call and its quantifier was incorrectly compiled,
461        leading to buffer overflow or other errors. This bug was discovered by the
462        LLVM fuzzer.
463    
464    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
465        assertion after (?(. The code was failing to check the character after
466        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
467        was discovered by the LLVM fuzzer.
468    
469    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
470        a fixed maximum following a group that contains a subroutine reference was
471        incorrectly compiled and could trigger buffer overflow. This bug was
472        discovered by the LLVM fuzzer.
473    
474    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
475        caused a stack overflow instead of the diagnosis of a non-fixed length
476        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
477    
478    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
479        (e.g. /(?<=\Ka)/) could make pcregrep loop.
480    
481    37. There was a similar problem to 36 in pcretest for global matches.
482    
483    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
484        and a subsequent item in the pattern caused a non-match, backtracking over
485        the repeated \X did not stop, but carried on past the start of the subject,
486        causing reference to random memory and/or a segfault. There were also some
487        other cases where backtracking after \C could crash. This set of bugs was
488        discovered by the LLVM fuzzer.
489    
490    39. The function for finding the minimum length of a matching string could take
491        a very long time if mutual recursion was present many times in a pattern,
492        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
493        method has been implemented. This infelicity was discovered by the LLVM
494        fuzzer.
495    
496    40. Static linking against the PCRE library using the pkg-config module was
497        failing on missing pthread symbols.
498    
499    
500    Version 8.36 26-September-2014
501    ------------------------------
502    
503    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
504        -Wmissing-field-initializers and -Wunused-parameter.
505    
506    2.  The tests for quantifiers being too big (greater than 65535) were being
507        applied after reading the number, and stupidly assuming that integer
508        overflow would give a negative number. The tests are now applied as the
509        numbers are read.
510    
511    3.  Tidy code in pcre_exec.c where two branches that used to be different are
512        now the same.
513    
514    4.  The JIT compiler did not generate match limit checks for certain
515        bracketed expressions with quantifiers. This may lead to exponential
516        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
517        issue should be resolved now.
518    
519    5.  Fixed an issue, which occures when nested alternatives are optimized
520        with table jumps.
521    
522    6.  Inserted two casts and changed some ints to size_t in the light of some
523        reported 64-bit compiler warnings (Bugzilla 1477).
524    
525    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
526        an empty string, which sometimes were behaving incorrectly in the
527        interpreter (though correctly in the JIT matcher). This pcretest input is
528        an example:
529    
530          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
531          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
532    
533        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
534        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
535        for an empty string was breaking the inner loop and carrying on at a lower
536        level, when possessive repeated groups should always return to a higher
537        level as they have no backtrack points in them. The empty string test now
538        occurs at the outer level.
539    
540    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
541        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
542    
543    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
544        doesn't).
545    
546    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
547        to hold a pattern to be incorrectly computed (too small) when there were
548        named back references to duplicated names. This could cause "internal
549        error: code overflow" or "double free or corruption" or other memory
550        handling errors.
551    
552    11. When named subpatterns had the same prefixes, back references could be
553        confused. For example, in this pattern:
554    
555          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
556    
557        the reference to 'Name' was incorrectly treated as a reference to a
558        duplicate name.
559    
560    12. A pattern such as /^s?c/mi8 where the optional character has more than
561        one "other case" was incorrectly compiled such that it would only try to
562        match starting at "c".
563    
564    13. When a pattern starting with \s was studied, VT was not included in the
565        list of possible starting characters; this should have been part of the
566        8.34/18 patch.
567    
568    14. If a character class started [\Qx]... where x is any character, the class
569        was incorrectly terminated at the ].
570    
571    15. If a pattern that started with a caseless match for a character with more
572        than one "other case" was studied, PCRE did not set up the starting code
573        unit bit map for the list of possible characters. Now it does. This is an
574        optimization improvement, not a bug fix.
575    
576    16. The Unicode data tables have been updated to Unicode 7.0.0.
577    
578    17. Fixed a number of memory leaks in pcregrep.
579    
580    18. Avoid a compiler warning (from some compilers) for a function call with
581        a cast that removes "const" from an lvalue by using an intermediate
582        variable (to which the compiler does not object).
583    
584    19. Incorrect code was compiled if a group that contained an internal recursive
585        back reference was optional (had quantifier with a minimum of zero). This
586        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
587        caused segmentation faults because of stack overflows at compile time.
588    
589    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
590        group that is quantified with an indefinite repeat, caused a compile-time
591        loop which used up all the system stack and provoked a segmentation fault.
592        This was not the same bug as 19 above.
593    
594    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
595        Patch by Mike Frysinger.
596    
597    
598    Version 8.35 04-April-2014
599    --------------------------
600    
601    1.  A new flag is set, when property checks are present in an XCLASS.
602        When this flag is not set, PCRE can perform certain optimizations
603        such as studying these XCLASS-es.
604    
605    2.  The auto-possessification of character sets were improved: a normal
606        and an extended character set can be compared now. Furthermore
607        the JIT compiler optimizes more character set checks.
608    
609    3.  Got rid of some compiler warnings for potentially uninitialized variables
610        that show up only when compiled with -O2.
611    
612    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
613        of a match later then the end of the match. The pcretest program was not
614        handling the case sensibly - it was outputting from the start to the next
615        binary zero. It now reports this situation in a message, and outputs the
616        text from the end to the start.
617    
618    5.  Fast forward search is improved in JIT. Instead of the first three
619        characters, any three characters with fixed position can be searched.
620        Search order: first, last, middle.
621    
622    6.  Improve character range checks in JIT. Characters are read by an inprecise
623        function now, which returns with an unknown value if the character code is
624        above a certain threshold (e.g: 256). The only limitation is that the value
625        must be bigger than the threshold as well. This function is useful when
626        the characters above the threshold are handled in the same way.
627    
628    7.  The macros whose names start with RAWUCHAR are placeholders for a future
629        mode in which only the bottom 21 bits of 32-bit data items are used. To
630        make this more memorable for those maintaining the code, the names have
631        been changed to start with UCHAR21, and an extensive comment has been added
632        to their definition.
633    
634    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
635        to the export list in Makefile.am (they were accidentally omitted from the
636        8.34 tarball).
637    
638    9.  The informational output from pcretest used the phrase "starting byte set"
639        which is inappropriate for the 16-bit and 32-bit libraries. As the output
640        for "first char" and "need char" really means "non-UTF-char", I've changed
641        "byte" to "char", and slightly reworded the output. The documentation about
642        these values has also been (I hope) clarified.
643    
644    10. Another JIT related optimization: use table jumps for selecting the correct
645        backtracking path, when more than four alternatives are present inside a
646        bracket.
647    
648    11. Empty match is not possible, when the minimum length is greater than zero,
649        and there is no \K in the pattern. JIT should avoid empty match checks in
650        such cases.
651    
652    12. In a caseless character class with UCP support, when a character with more
653        than one alternative case was not the first character of a range, not all
654        the alternative cases were added to the class. For example, s and \x{17f}
655        are both alternative cases for S: the class [RST] was handled correctly,
656        but [R-T] was not.
657    
658    13. The configure.ac file always checked for pthread support when JIT was
659        enabled. This is not used in Windows, so I have put this test inside a
660        check for the presence of windows.h (which was already tested for).
661    
662    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
663        The algorithm provides a way to skip certain starting offsets, and usually
664        faster than linear prefix searches.
665    
666    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
667        as for 'fr_FR' and 'french'. For some reason, however, it then used the
668        Windows-specific input and output files, which have 'french' screwed in.
669        So this could never have worked. One of the problems with locales is that
670        they aren't always the same. I have now updated RunTest so that it checks
671        the output of the locale test (test 3) against three different output
672        files, and it allows the test to pass if any one of them matches. With luck
673        this should make the test pass on some versions of Solaris where it was
674        failing. Because of the uncertainty, the script did not used to stop if
675        test 3 failed; it now does. If further versions of a French locale ever
676        come to light, they can now easily be added.
677    
678    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
679        there was a message during ./configure, but it did not stop. This now
680        provokes an error. The invalid example in README has been corrected.
681        If a value less than the minimum is given, the minimum value has always
682        been used, but now a warning is given.
683    
684    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
685        was a bug in the test system, which is now fixed. Also, the list of various
686        configurations that are tested for each release did not have one with both
687        16/32 bits and --enable-bar-anycrlf. It now does.
688    
689    18. pcretest was missing "-C bsr" for displaying the \R default setting.
690    
691    19. Little endian PowerPC systems are supported now by the JIT compiler.
692    
693    20. The fast forward newline mechanism could enter to an infinite loop on
694        certain invalid UTF-8 input. Although we don't support these cases
695        this issue can be fixed by a performance optimization.
696    
697    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
698        not take account if existing stack usage. There is now a new global
699        variable called pcre_stack_guard that can be set to point to an external
700        function to check stack availability. It is called at the start of
701        processing every parenthesized group.
702    
703    22. A typo in the code meant that in ungreedy mode the max/min qualifier
704        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
705        not match "ab".
706    
707    23. When UTF was disabled, the JIT program reported some incorrect compile
708        errors. These messages are silenced now.
709    
710    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
711        compiler.
712    
713    25. Change all the temporary files used in RunGrepTest to be different to those
714        used by RunTest so that the tests can be run simultaneously, for example by
715        "make -j check".
716    
717    
718    Version 8.34 15-December-2013
719    -----------------------------
720    
721    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
722        executable memory. Patch inspired by Carsten Klein.
723    
724    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
725        this macro is never tested and has no effect, because the work to support
726        coverage involves only compiling and linking options and special targets in
727        the Makefile. The comment in config.h implied that defining the macro would
728        enable coverage support, which is totally false. There was also support for
729        setting this macro in the CMake files (my fault, I just copied it from
730        configure). SUPPORT_GCOV has now been removed.
731    
732    3.  Make a small performance improvement in strlen16() and strlen32() in
733        pcretest.
734    
735    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
736        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
737        these cases). There was also one in pcretest.c.
738    
739    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
740    
741    6.  In UTF mode, the code for checking whether a group could match an empty
742        string (which is used for indefinitely repeated groups to allow for
743        breaking an infinite loop) was broken when the group contained a repeated
744        negated single-character class with a character that occupied more than one
745        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
746        UTF-8 mode). The effect was undefined: the group might or might not be
747        deemed as matching an empty string, or the program might have crashed.
748    
749    7.  The code for checking whether a group could match an empty string was not
750        recognizing that \h, \H, \v, \V, and \R must match a character.
751    
752    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
753        an empty string. If it can, pcretest shows this in its information output.
754    
755    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
756        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
757        matched by pcre_exec() without using JIT:
758    
759        (a) If the rest of the pattern did not match after a maximal run of
760            grapheme clusters, the code for backing up to try with fewer of them
761            did not always back up over a full grapheme when characters that do not
762            have the modifier quality were involved, e.g. Hangul syllables.
763    
764        (b) If the match point in a subject started with modifier character, and
765            there was no match, the code could incorrectly back up beyond the match
766            point, and potentially beyond the first character in the subject,
767            leading to a segfault or an incorrect match result.
768    
769    10. A conditional group with an assertion condition could lead to PCRE
770        recording an incorrect first data item for a match if no other first data
771        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
772        first data item, and therefore matched "ca" after "c" instead of at the
773        start.
774    
775    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
776        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
777    
778    12. The source of pcregrep now includes z/OS-specific code so that it can be
779        compiled for z/OS as part of the special z/OS distribution.
780    
781    13. Added the -T and -TM options to pcretest.
782    
783    14. The code in pcre_compile.c for creating the table of named capturing groups
784        has been refactored. Instead of creating the table dynamically during the
785        actual compiling pass, the information is remembered during the pre-compile
786        pass (on the stack unless there are more than 20 named groups, in which
787        case malloc() is used) and the whole table is created before the actual
788        compile happens. This has simplified the code (it is now nearly 150 lines
789        shorter) and prepared the way for better handling of references to groups
790        with duplicate names.
791    
792    15. A back reference to a named subpattern when there is more than one of the
793        same name now checks them in the order in which they appear in the pattern.
794        The first one that is set is used for the reference. Previously only the
795        first one was inspected. This change makes PCRE more compatible with Perl.
796    
797    16. Unicode character properties were updated from Unicode 6.3.0.
798    
799    17. The compile-time code for auto-possessification has been refactored, based
800        on a patch by Zoltan Herczeg. It now happens after instead of during
801        compilation. The code is cleaner, and more cases are handled. The option
802        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
803        options in pcretest are provided to set it. It can also be set by
804        (*NO_AUTO_POSSESS) at the start of a pattern.
805    
806    18. The character VT has been added to the default ("C" locale) set of
807        characters that match \s and are generally treated as white space,
808        following this same change in Perl 5.18. There is now no difference between
809        "Perl space" and "POSIX space". Whether VT is treated as white space in
810        other locales depends on the locale.
811    
812    19. The code for checking named groups as conditions, either for being set or
813        for being recursed, has been refactored (this is related to 14 and 15
814        above). Processing unduplicated named groups should now be as fast at
815        numerical groups, and processing duplicated groups should be faster than
816        before.
817    
818    20. Two patches to the CMake build system, by Alexander Barkov:
819    
820          (1) Replace the "source" command by "." in CMakeLists.txt because
821              "source" is a bash-ism.
822    
823          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
824              without these the CMake build does not work on Solaris.
825    
826    21. Perl has changed its handling of \8 and \9. If there is no previously
827        encountered capturing group of those numbers, they are treated as the
828        literal characters 8 and 9 instead of a binary zero followed by the
829        literals. PCRE now does the same.
830    
831    22. Following Perl, added \o{} to specify codepoints in octal, making it
832        possible to specify values greater than 0777 and also making them
833        unambiguous.
834    
835    23. Perl now gives an error for missing closing braces after \x{... instead of
836        treating the string as literal. PCRE now does the same.
837    
838    24. RunTest used to grumble if an inappropriate test was selected explicitly,
839        but just skip it when running all tests. This make it awkward to run ranges
840        of tests when one of them was inappropriate. Now it just skips any
841        inappropriate tests, as it always did when running all tests.
842    
843    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
844        character types such as \d or \w, too many callouts were inserted, and the
845        data that they returned was rubbish.
846    
847    26. In UCP mode, \s was not matching two of the characters that Perl matches,
848        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
849        were matched by \h. The code has now been refactored so that the lists of
850        the horizontal and vertical whitespace characters used for \h and \v (which
851        are defined only in one place) are now also used for \s.
852    
853    27. Add JIT support for the 64 bit TileGX architecture.
854        Patch by Jiong Wang (Tilera Corporation).
855    
856    28. Possessive quantifiers for classes (both explicit and automatically
857        generated) now use special opcodes instead of wrapping in ONCE brackets.
858    
859    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
860        (because it's meaningless), this was not happening when PCRE_CASELESS was
861        set. Not wrong, but inefficient.
862    
863    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
864        properties for \w, \d, etc) is present in a test regex. Otherwise if the
865        test contains no characters greater than 255, Perl doesn't realise it
866        should be using Unicode semantics.
867    
868    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
869        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
870        does in Unicode mode.
871    
872    32. Added the "forbid" facility to pcretest so that putting tests into the
873        wrong test files can sometimes be quickly detected.
874    
875    33. There is now a limit (default 250) on the depth of nesting of parentheses.
876        This limit is imposed to control the amount of system stack used at compile
877        time. It can be changed at build time by --with-parens-nest-limit=xxx or
878        the equivalent in CMake.
879    
880    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
881        errors. Perl warns for these when in warning mode, but PCRE has no facility
882        for giving warnings.
883    
884    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
885        However, this was not working for (?!) because it is optimized to (*FAIL),
886        for which PCRE does not allow quantifiers. The optimization is now disabled
887        when a quantifier follows (?!). I can't see any use for this, but it makes
888        things uniform.
889    
890    36. Perl no longer allows group names to start with digits, so I have made this
891        change also in PCRE. It simplifies the code a bit.
892    
893    37. In extended mode, Perl ignores spaces before a + that indicates a
894        possessive quantifier. PCRE allowed a space before the quantifier, but not
895        before the possessive +. It now does.
896    
897    38. The use of \K (reset reported match start) within a repeated possessive
898        group such as (a\Kb)*+ was not working.
899    
900    40. Document that the same character tables must be used at compile time and
901        run time, and that the facility to pass tables to pcre_exec() and
902        pcre_dfa_exec() is for use only with saved/restored patterns.
903    
904    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
905        features for Builds with MSVC:
906    
907        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
908           builds). The .rc files can be used to set FileDescription and many other
909           attributes.
910    
911        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
912           This allows higher-level build scripts which want .pdb files to avoid
913           hard-coding the exact files needed."
914    
915    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
916        mean "start of word" and "end of word", respectively, as a transition aid.
917    
918    43. A minimizing repeat of a class containing codepoints greater than 255 in
919        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
920        compiled to use the heap for recursion.
921    
922    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
923        is configured.
924    
925    
926    Version 8.33 28-May-2013
927    ------------------------
928    
929    1.  Added 'U' to some constants that are compared to unsigned integers, to
930        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
931        variables that are added to signed variables, to ensure the result is
932        signed and can be negated.
933    
934    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
935        CMake config files.
936    
937    3.  Revise the creation of config.h.generic so that all boolean macros are
938        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
939        overriding via -D on the command line possible.
940    
941    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
942        to unsigned int is reported to make a quite noticeable speed difference in
943        a specific Windows environment. Testing on Linux did also appear to show
944        some benefit (and it is clearly not harmful). Also fixed the definition of
945        Xop which should be unsigned.
946    
947    5.  Related to (4), changing the definition of the intermediate variable cc
948        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
949        improvements.
950    
951    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
952        unnecessary spaces.
953    
954    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
955        and later.
956    
957    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
958    
959    9.  Optimizing fast_forward_start_bits in JIT.
960    
961    10. Adding support for callouts in JIT, and fixing some issues revealed
962        during this work. Namely:
963    
964        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
965    
966        (b) Minimum length was not checked before the matching is started.
967    
968    11. The value of capture_last that is passed to callouts was incorrect in some
969        cases when there was a capture on one path that was subsequently abandoned
970        after a backtrack. Also, the capture_last value is now reset after a
971        recursion, since all captures are also reset in this case.
972    
973    12. The interpreter no longer returns the "too many substrings" error in the
974        case when an overflowing capture is in a branch that is subsequently
975        abandoned after a backtrack.
976    
977    13. In the pathological case when an offset vector of size 2 is used, pcretest
978        now prints out the matched string after a yield of 0 or 1.
979    
980    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
981        Only supported by the JIT compiler at the moment.
982    
983    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
984    
985    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
986        offset of the starting point of the matching process, provided the offsets
987        vector is large enough.
988    
989    17. The \A escape now records a lookbehind value of 1, though its execution
990        does not actually inspect the previous character. This is to ensure that,
991        in partial multi-segment matching, at least one character from the old
992        segment is retained when a new segment is processed. Otherwise, if there
993        are no lookbehinds in the pattern, \A might match incorrectly at the start
994        of a new segment.
995    
996    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
997    
998    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
999        gives some modest performance improvement in 8-bit mode.
1000    
1001    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1002        be expressed in certain programming languages using Universal Character
1003        Names.
1004    
1005    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1006        which points out that "non characters" are not "characters that may not
1007        appear in Unicode strings" but rather "characters that are reserved for
1008        internal use and have only local meaning".
1009    
1010    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1011        there was a conditional group that depended on an assertion, if the
1012        assertion was false, the callout that immediately followed the alternation
1013        in the condition was skipped when pcre_exec() was used for matching.
1014    
1015    23. Allow an explicit callout to be inserted before an assertion that is the
1016        condition for a conditional group, for compatibility with automatic
1017        callouts, which always insert a callout at this point.
1018    
1019    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1020        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1021    
1022    25. (*PRUNE) is now supported by the JIT compiler.
1023    
1024    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1025    
1026    27. Fix the case where there are two or more SKIPs with arguments that may be
1027        ignored.
1028    
1029    28. (*SKIP) is now supported by the JIT compiler.
1030    
1031    29. (*THEN) is now supported by the JIT compiler.
1032    
1033    30. Update RunTest with additional test selector options.
1034    
1035    31. The way PCRE handles backtracking verbs has been changed in two ways.
1036    
1037        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1038        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1039        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1040        rather obscure rules do not always do the same thing.
1041    
1042        (2) Previously, backtracking verbs were confined within assertions. This is
1043        no longer the case for positive assertions, except for (*ACCEPT). Again,
1044        this sometimes improves Perl compatibility, and sometimes does not.
1045    
1046    32. A number of tests that were in test 2 because Perl did things differently
1047        have been moved to test 1, because either Perl or PCRE has changed, and
1048        these tests are now compatible.
1049    
1050    32. Backtracking control verbs are now handled in the same way in JIT and
1051        interpreter.
1052    
1053    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1054        contained a forward subroutine reference caused a compile error.
1055    
1056    34. Auto-detect and optimize limited repetitions in JIT.
1057    
1058    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1059        blocking (*UTF) etc.
1060    
1061    36. In the interpreter, maximizing pattern repetitions for characters and
1062        character types now use tail recursion, which reduces stack usage.
1063    
1064    37. The value of the max lookbehind was not correctly preserved if a compiled
1065        and saved regex was reloaded on a host of different endianness.
1066    
1067    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1068        of the compiled pattern block, expand the flags field from 16 to 32 bits
1069        because it was almost full.
1070    
1071    39. Try madvise first before posix_madvise.
1072    
1073    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1074        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1075        original change locked out all matches of empty strings. This has been
1076        changed so that one match of an empty string per line is recognized.
1077        Subsequent searches on the same line (for colouring or for --only-matching,
1078        for example) do not recognize empty strings.
1079    
1080    41. Applied a user patch to fix a number of spelling mistakes in comments.
1081    
1082    42. Data lines longer than 65536 caused pcretest to crash.
1083    
1084    43. Clarified the data type for length and startoffset arguments for pcre_exec
1085        and pcre_dfa_exec in the function-specific man pages, where they were
1086        explicitly stated to be in bytes, never having been updated. I also added
1087        some clarification to the pcreapi man page.
1088    
1089    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1090        a segmentation fault.
1091    
1092    
1093    Version 8.32 30-November-2012
1094    -----------------------------
1095    
1096    1.  Improved JIT compiler optimizations for first character search and single
1097        character iterators.
1098    
1099    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1100        Patch by Daniel Richard G.
1101    
1102    3.  Single character iterator optimizations in the JIT compiler.
1103    
1104    4.  Improved JIT compiler optimizations for character ranges.
1105    
1106    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1107        Reported by Giuseppe D'Angelo.
1108    
1109    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1110        of a line, was being set incorrectly in cases where .* appeared inside
1111        atomic brackets at the start of a pattern, or where there was a subsequent
1112        *PRUNE or *SKIP.
1113    
1114    7.  Improved instruction cache flush for POWER/PowerPC.
1115        Patch by Daniel Richard G.
1116    
1117    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1118        grep:
1119    
1120        (a) There is now no limit to the number of patterns to be matched.
1121    
1122        (b) An error is given if a pattern is too long.
1123    
1124        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1125            are now supported.
1126    
1127        (d) --exclude-from and --include-from (multiple use) have been added.
1128    
1129        (e) Exclusions and inclusions now apply to all files and directories, not
1130            just to those obtained from scanning a directory recursively.
1131    
1132        (f) Multiple uses of -f and --file-list are now supported.
1133    
1134        (g) In a Windows environment, the default for -d has been changed from
1135            "read" (the GNU grep default) to "skip", because otherwise the presence
1136            of a directory in the file list provokes an error.
1137    
1138        (h) The documentation has been revised and clarified in places.
1139    
1140    9.  Improve the matching speed of capturing brackets.
1141    
1142    10. Changed the meaning of \X so that it now matches a Unicode extended
1143        grapheme cluster.
1144    
1145    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1146        out POSIX threads when JIT support is configured.
1147    
1148    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1149    
1150    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1151        structure could go wrong in environments where size_t is not the same size
1152        as int.
1153    
1154    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1155        set.
1156    
1157    15. The EBCDIC support had decayed; later updates to the code had included
1158        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1159        general tidy up of EBCDIC-related issues, and the documentation was also
1160        not quite right. There is now a test that can be run on ASCII systems to
1161        check some of the EBCDIC-related things (but is it not a full test).
1162    
1163    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1164        in a small tidy to the code.
1165    
1166    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1167    
1168    18. If the --only-matching (-o) option in pcregrep is specified multiple
1169        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1170        substrings matched by the 1st and 2nd capturing parentheses. A separating
1171        string can be specified by --om-separator (default empty).
1172    
1173    19. Improving the first n character searches.
1174    
1175    20. Turn case lists for horizontal and vertical white space into macros so that
1176        they are defined only once.
1177    
1178    21. This set of changes together give more compatible Unicode case-folding
1179        behaviour for characters that have more than one other case when UCP
1180        support is available.
1181    
1182        (a) The Unicode property table now has offsets into a new table of sets of
1183            three or more characters that are case-equivalent. The MultiStage2.py
1184            script that generates these tables (the pcre_ucd.c file) now scans
1185            CaseFolding.txt instead of UnicodeData.txt for character case
1186            information.
1187    
1188        (b) The code for adding characters or ranges of characters to a character
1189            class has been abstracted into a generalized function that also handles
1190            case-independence. In UTF-mode with UCP support, this uses the new data
1191            to handle characters with more than one other case.
1192    
1193        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1194            whose other case is greater than 256 are now correctly matched
1195            caselessly. Previously, the high codepoint matched the low one, but not
1196            vice versa.
1197    
1198        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1199            of the new class addition function, using character lists defined as
1200            macros alongside the case definitions of 20 above.
1201    
1202        (e) Caseless back references now work with characters that have more than
1203            one other case.
1204    
1205        (f) General caseless matching of characters with more than one other case
1206            is supported.
1207    
1208    22. Unicode character properties were updated from Unicode 6.2.0
1209    
1210    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1211    
1212    24. Add support for 32-bit character strings, and UTF-32
1213    
1214    25. Major JIT compiler update (code refactoring and bugfixing).
1215        Experimental Sparc 32 support is added.
1216    
1217    26. Applied a modified version of Daniel Richard G's patch to create
1218        pcre.h.generic and config.h.generic by "make" instead of in the
1219        PrepareRelease script.
1220    
1221    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1222        pcre_compile.c when checking for a zero character.
1223    
1224    28. Introducing a native interface for JIT. Through this interface, the compiled
1225        machine code can be directly executed. The purpose of this interface is to
1226        provide fast pattern matching, so several sanity checks are not performed.
1227        However, feature tests are still performed. The new interface provides
1228        1.4x speedup compared to the old one.
1229    
1230    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1231        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1232        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1233    
1234    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1235        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1236        the "old" RFC 2279). Instead, it ended up passing a negative length to
1237        pcre_exec().
1238    
1239    31. Add support for GCC's visibility feature to hide internal functions.
1240    
1241    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1242        "unknown -C option" after outputting 0 or 1.
1243    
1244    33. There is now support for generating a code coverage report for the test
1245        suite in environments where gcc is the compiler and lcov is installed. This
1246        is mainly for the benefit of the developers.
1247    
1248    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1249        unaddressable using valgrind annotations, allowing valgrind to detect
1250        invalid memory accesses. This is mainly for the benefit of the developers.
1251    
1252    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1253    
1254    26. Give configure error if --enable-cpp but no C++ compiler found.
1255    
1256    
1257    Version 8.31 06-July-2012
1258    -------------------------
1259    
1260    1.  Fixing a wrong JIT test case and some compiler warnings.
1261    
1262    2.  Removed a bashism from the RunTest script.
1263    
1264    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1265        to unsigned type, result still unsigned" that was given by an MS compiler
1266        on encountering the code "-sizeof(xxx)".
1267    
1268    4.  Partial matching support is added to the JIT compiler.
1269    
1270    5.  Fixed several bugs concerned with partial matching of items that consist
1271        of more than one character:
1272    
1273        (a) /^(..)\1/ did not partially match "aba" because checking references was
1274            done on an "all or nothing" basis. This also applied to repeated
1275            references.
1276    
1277        (b) \R did not give a hard partial match if \r was found at the end of the
1278            subject.
1279    
1280        (c) \X did not give a hard partial match after matching one or more
1281            characters at the end of the subject.
1282    
1283        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1284            a partial match for the string "\r".
1285    
1286        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1287            a partial match for a CR character at the end of the subject string.
1288    
1289    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1290        running pcretest, the text "(JIT)" added to the output whenever JIT is
1291        actually used to run the match.
1292    
1293    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1294        or /S+[+] with a digit between 1 and 7.
1295    
1296    8.  OP_NOT now supports any UTF character not just single-byte ones.
1297    
1298    9.  (*MARK) control verb is now supported by the JIT compiler.
1299    
1300    10. The command "./RunTest list" lists the available tests without actually
1301        running any of them. (Because I keep forgetting what they all are.)
1302    
1303    11. Add PCRE_INFO_MAXLOOKBEHIND.
1304    
1305    12. Applied a (slightly modified) user-supplied patch that improves performance
1306        when the heap is used for recursion (compiled with --disable-stack-for-
1307        recursion). Instead of malloc and free for each heap frame each time a
1308        logical recursion happens, frames are retained on a chain and re-used where
1309        possible. This sometimes gives as much as 30% improvement.
1310    
1311    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1312        call.
1313    
1314    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1315    
1316    15. It is now possible to link pcretest with libedit as an alternative to
1317        libreadline.
1318    
1319    16. (*COMMIT) control verb is now supported by the JIT compiler.
1320    
1321    17. The Unicode data tables have been updated to Unicode 6.1.0.
1322    
1323    18. Added --file-list option to pcregrep.
1324    
1325    19. Added binary file support to pcregrep, including the -a, --binary-files,
1326        -I, and --text options.
1327    
1328    20. The madvise function is renamed for posix_madvise for QNX compatibility
1329        reasons. Fixed by Giuseppe D'Angelo.
1330    
1331    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1332    
1333    22. Changed the option for creating symbolic links for 16-bit man pages from
1334        -s to -sf so that re-installing does not cause issues.
1335    
1336    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1337    
1338    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1339        matches in certain environments (the workspace was not being correctly
1340        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1341        some of the workspace data at the beginning of a restart.
1342    
1343    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1344        was not doing so when it should - probably a typo introduced by SVN 528
1345        (change 8.10/14).
1346    
1347    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1348        \w+ when the character tables indicated that \x{c4} was a word character.
1349        There were several related cases, all because the tests for doing a table
1350        lookup were testing for characters less than 127 instead of 255.
1351    
1352    27. If a pattern contains capturing parentheses that are not used in a match,
1353        their slots in the ovector are set to -1. For those that are higher than
1354        any matched groups, this happens at the end of processing. In the case when
1355        there were back references that the ovector was too small to contain
1356        (causing temporary malloc'd memory to be used during matching), and the
1357        highest capturing number was not used, memory off the end of the ovector
1358        was incorrectly being set to -1. (It was using the size of the temporary
1359        memory instead of the true size.)
1360    
1361    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1362        ovector size, it uses memory at the end of the block that it has got.
1363    
1364    29. Check for an overlong MARK name and give an error at compile time. The
1365        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1366    
1367    30. JIT compiler update.
1368    
1369    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1370        Rill for the patch.
1371    
1372    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1373    
1374    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1375    
1376    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1377        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1378        was enabled.
1379    
1380    35. Improve JIT code generation for greedy plus quantifier.
1381    
1382    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1383        1 to "aa" instead of to an empty string. The bug affected repeated groups
1384        that could potentially match an empty string.
1385    
1386    37. Optimizing single character iterators in JIT.
1387    
1388    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1389        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1390        codepoints that are too big for the mode are faulted, and in a UTF mode,
1391        disallowed codepoints are also faulted.
1392    
1393    39. If PCRE was compiled with UTF support, in three places in the DFA
1394        matcher there was code that should only have been obeyed in UTF mode, but
1395        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1396        processing when bytes with values greater than 127 were present. In 16-bit
1397        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1398        both cases the values are those that cannot be the first data item in a UTF
1399        character. The three items that might have provoked this were recursions,
1400        possessively repeated groups, and atomic groups.
1401    
1402    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1403        and pcregrep, because some OS require shared objects to be explicitly
1404        passed to ld, causing the link step to fail if they are not.
1405    
1406    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1407        mode, patterns that started with \h* or \R* might be incorrectly matched.
1408    
1409    
1410    Version 8.30 04-February-2012
1411    -----------------------------
1412    
1413    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1414        name is defined in ctype.h.
1415    
1416    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1417        only in quite long subpatterns.
1418    
1419    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1420        since it was replaced by pcre_fullinfo() in February 2000.
1421    
1422    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1423        match a (*MARK), and the match failed at the start of the subject, a
1424        reference to memory before the start of the subject could occur. This bug
1425        was introduced by fix 17 of release 8.21.
1426    
1427    5.  A reference to an unset group with zero minimum repetition was giving
1428        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1429        /(another)?(\1?)test/ matched against "hello world test". This bug was
1430        introduced in release 8.13.
1431    
1432    6.  Add support for 16-bit character strings (a large amount of work involving
1433        many changes and refactorings).
1434    
1435    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1436        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1437        from a file.
1438    
1439    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1440        rounding is not applied in this particular case).
1441    
1442    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1443        if they appear, or are escaped, in patterns.
1444    
1445    10. Get rid of a number of -Wunused-but-set-variable warnings.
1446    
1447    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1448        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1449        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1450        also returns the mark "x". This bug applied to capturing parentheses,
1451        non-capturing parentheses, and atomic parentheses. It also applied to some
1452        assertions.
1453    
1454    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1455        information out of configure.ac instead of relying on pcre.h.generic, which
1456        is not stored in the repository.
1457    
1458    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1459        -lreadline.
1460    
1461    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1462    
1463    15. Applied Graycode's patch to put the top-level frame on the stack rather
1464        than the heap when not using the stack for recursion. This gives a
1465        performance improvement in many cases when recursion is not deep.
1466    
1467    16. Experimental code added to "pcretest -C" to output the stack frame size.
1468    
1469    
1470    Version 8.21 12-Dec-2011
1471    ------------------------
1472    
1473    1.  Updating the JIT compiler.
1474    
1475    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1476        are added as well.
1477    
1478    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1479        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1480        calling _pcre_jit_exec. Some extra comments are added.
1481    
1482    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1483        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1484        was introduced by change 18 for 8.20.
1485    
1486    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1487        ECMA-262 standard.
1488    
1489    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1490        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1491        This bug was probably introduced by change 9 of 8.13.
1492    
1493    7.  While fixing 6 above, I noticed that a number of other items were being
1494        incorrectly rejected as "not fixed length". This arose partly because newer
1495        opcodes had not been added to the fixed-length checking code. I have (a)
1496        corrected the bug and added tests for these items, and (b) arranged for an
1497        error to occur if an unknown opcode is encountered while checking for fixed
1498        length instead of just assuming "not fixed length". The items that were
1499        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1500        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1501        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1502    
1503    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1504        being incorrectly compiled and would have given unpredicatble results.
1505    
1506    9.  A possessively repeated subpattern with minimum repeat count greater than
1507        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1508        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1509        the first (A) could occur when it should not.
1510    
1511    10. Add a cast and remove a redundant test from the code.
1512    
1513    11. JIT should use pcre_malloc/pcre_free for allocation.
1514    
1515    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1516        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1517        is anything other than /usr, -L is still shown).
1518    
1519    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1520    
1521    14. Perl does not support \N without a following name in a [] class; PCRE now
1522        also gives an error.
1523    
1524    15. If a forward reference was repeated with an upper limit of around 2000,
1525        it caused the error "internal error: overran compiling workspace". The
1526        maximum number of forward references (including repeats) was limited by the
1527        internal workspace, and dependent on the LINK_SIZE. The code has been
1528        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1529        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1530        of around 200,000 forward references. While doing this, I also speeded up
1531        the filling in of repeated forward references.
1532    
1533    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1534        incorrectly expecting the subject to contain another "a" after the start.
1535    
1536    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1537        in the match, the SKIP should be ignored. This was not happening; instead
1538        the SKIP was being treated as NOMATCH. For patterns such as
1539        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1540        tested.
1541    
1542    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1543        now much more compatible with Perl, in particular in cases where the result
1544        is a non-match for a non-anchored pattern. For example, if
1545        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1546        "m", where previously it did not return a name. A side effect of this
1547        change is that for partial matches, the last encountered mark name is
1548        returned, as for non matches. A number of tests that were previously not
1549        Perl-compatible have been moved into the Perl-compatible test files. The
1550        refactoring has had the pleasing side effect of removing one argument from
1551        the match() function, thus reducing its stack requirements.
1552    
1553    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1554        subsequent uses of /S (without +) incorrectly behaved like /S+.
1555    
1556    21. Retrieve executable code size support for the JIT compiler and fixing
1557        some warnings.
1558    
1559    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1560        not work when the shorter character appeared right at the end of the
1561        subject string.
1562    
1563    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1564        systems.
1565    
1566    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1567        output it when the /M option is used in pcretest.
1568    
1569    25. The CheckMan script was not being included in the distribution. Also, added
1570        an explicit "perl" to run Perl scripts from the PrepareRelease script
1571        because this is reportedly needed in Windows.
1572    
1573    26. If study data was being save in a file and studying had not found a set of
1574        "starts with" bytes for the pattern, the data written to the file (though
1575        never used) was taken from uninitialized memory and so caused valgrind to
1576        complain.
1577    
1578    27. Updated RunTest.bat as provided by Sheri Pierce.
1579    
1580    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1581    
1582    29. Computation of memory usage for the table of capturing group names was
1583        giving an unnecessarily large value.
1584    
1585    
1586    Version 8.20 21-Oct-2011
1587    ------------------------
1588    
1589    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1590        a POSIX class. After further experiments with Perl, which convinced me that
1591        Perl has bugs and confusions, a closing square bracket is no longer allowed
1592        in a POSIX name. This bug also affected patterns with classes that started
1593        with full stops.
1594    
1595    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1596        captured substring, but while checking the failing first alternative,
1597        substring 1 is temporarily captured. If the output vector supplied to
1598        pcre_exec() was not big enough for this capture, the yield of the function
1599        was still zero ("insufficient space for captured substrings"). This cannot
1600        be totally fixed without adding another stack variable, which seems a lot
1601        of expense for a edge case. However, I have improved the situation in cases
1602        such as /(a)(b)x|abc/ matched against "abc", where the return code
1603        indicates that fewer than the maximum number of slots in the ovector have
1604        been set.
1605    
1606    3.  Related to (2) above: when there are more back references in a pattern than
1607        slots in the output vector, pcre_exec() uses temporary memory during
1608        matching, and copies in the captures as far as possible afterwards. It was
1609        using the entire output vector, but this conflicts with the specification
1610        that only 2/3 is used for passing back captured substrings. Now it uses
1611        only the first 2/3, for compatibility. This is, of course, another edge
1612        case.
1613    
1614    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1615        main code base, and can be used by building with --enable-jit. When this is
1616        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1617        runtime --no-jit option is given.
1618    
1619    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1620        ovector, the return from the function was zero, implying that there were
1621        other matches that did not fit. The correct "exactly full" value is now
1622        returned.
1623    
1624    6.  If a subpattern that was called recursively or as a subroutine contained
1625        (*PRUNE) or any other control that caused it to give a non-standard return,
1626        invalid errors such as "Error -26 (nested recursion at the same subject
1627        position)" or even infinite loops could occur.
1628    
1629    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1630        computing the minimum length on reaching *ACCEPT, and so ended up with the
1631        wrong value of 1 rather than 0. Further investigation indicates that
1632        computing a minimum subject length in the presence of *ACCEPT is difficult
1633        (think back references, subroutine calls), and so I have changed the code
1634        so that no minimum is registered for a pattern that contains *ACCEPT.
1635    
1636    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1637        it was not handled as intended. [But see 16 below.]
1638    
1639    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1640        Sheri Pierce.
1641    
1642    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1643        the first byte in a match must be "a".
1644    
1645    11. Change 17 for 8.13 increased the recursion depth for patterns like
1646        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1647        pattern contains any instances of (*THEN). If it does not, the old
1648        optimizations are restored. It would be nice to do this on a per-group
1649        basis, but at the moment that is not feasible.
1650    
1651    12. In some environments, the output of pcretest -C is CRLF terminated. This
1652        broke RunTest's code that checks for the link size. A single white space
1653        character after the value is now allowed for.
1654    
1655    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1656        For "fr", it uses the Windows-specific input and output files.
1657    
1658    14. If (*THEN) appeared in a group that was called recursively or as a
1659        subroutine, it did not work as intended. [But see next item.]
1660    
1661    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1662        pattern fragments (but not containing any | characters). If A and B are
1663        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1664        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1665        D. In other words, Perl considers parentheses that do not contain any |
1666        characters to be part of a surrounding alternative, whereas PCRE was
1667        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1668        differently. PCRE now behaves in the same way as Perl, except in the case
1669        of subroutine/recursion calls such as (?1) which have in any case always
1670        been different (but PCRE had them first :-).
1671    
1672    16. Related to 15 above: Perl does not treat the | in a conditional group as
1673        creating alternatives. Such a group is treated in the same way as an
1674        ordinary group without any | characters when processing (*THEN). PCRE has
1675        been changed to match Perl's behaviour.
1676    
1677    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1678        RunGrepTest script failed.
1679    
1680    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1681        inevitable for groups that contain captures, but it can lead to a lot of
1682        stack use in large patterns. The old behaviour has been restored for atomic
1683        groups that do not contain any capturing parentheses.
1684    
1685    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1686        suppress the check for a minimum subject length at run time. (If it was
1687        given to pcre_exec() or pcre_dfa_exec() it did work.)
1688    
1689    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1690        fail to work when decoding hex characters in data strings in EBCDIC
1691        environments.
1692    
1693    21. It appears that in at least one Mac OS environment, the isxdigit() function
1694        is implemented as a macro that evaluates to its argument more than once,
1695        contravening the C 90 Standard (I haven't checked a later standard). There
1696        was an instance in pcretest which caused it to go wrong when processing
1697        \x{...} escapes in subject strings. The has been rewritten to avoid using
1698        things like p++ in the argument of isxdigit().
1699    
1700    
1701    Version 8.13 16-Aug-2011
1702    ------------------------
1703    
1704    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1705    
1706    2.  Two minor typos in pcre_internal.h have been fixed.
1707    
1708    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1709        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1710        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1711    
1712    4.  There were a number of related bugs in the code for matching backrefences
1713        caselessly in UTF-8 mode when codes for the characters concerned were
1714        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1715        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1716        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1717        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1718        2-byte code at the end of the subject (it thought there wasn't enough data
1719        left).
1720    
1721    5.  Comprehensive information about what went wrong is now returned by
1722        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1723        as the output vector has at least 2 elements. The offset of the start of
1724        the failing character and a reason code are placed in the vector.
1725    
1726    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1727        now returned is for the first byte of the failing character, instead of the
1728        last byte inspected. This is an incompatible change, but I hope it is small
1729        enough not to be a problem. It makes the returned offset consistent with
1730        pcre_exec() and pcre_dfa_exec().
1731    
1732    7.  pcretest now gives a text phrase as well as the error number when
1733        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1734        failure, the offset and reason code are output.
1735    
1736    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1737        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1738        back over a single character (\n). This seems wrong (because it treated the
1739        two characters as a single entity when going forwards), conflicts with the
1740        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1741        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1742        has been changed.
1743    
1744    9.  Some internal refactoring has changed the processing so that the handling
1745        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1746        time (the PCRE_DOTALL option was changed this way some time ago: version
1747        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1748        which was always a bit of a fudge. It also means that there is one less
1749        argument for the match() function, which reduces its stack requirements
1750        slightly. This change also fixes an incompatibility with Perl: the pattern
1751        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1752    
1753    10. More internal refactoring has drastically reduced the number of recursive
1754        calls to match() for possessively repeated groups such as (abc)++ when
1755        using pcre_exec().
1756    
1757    11. While implementing 10, a number of bugs in the handling of groups were
1758        discovered and fixed:
1759    
1760        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1761        (a|)*(?1) gave a compile-time internal error.
1762        ((a|)+)+  did not notice that the outer group could match an empty string.
1763        (^a|^)+   was not marked as anchored.
1764        (.*a|.*)+ was not marked as matching at start or after a newline.
1765    
1766    12. Yet more internal refactoring has removed another argument from the match()
1767        function. Special calls to this function are now indicated by setting a
1768        value in a variable in the "match data" data block.
1769    
1770    13. Be more explicit in pcre_study() instead of relying on "default" for
1771        opcodes that mean there is no starting character; this means that when new
1772        ones are added and accidentally left out of pcre_study(), testing should
1773        pick them up.
1774    
1775    14. The -s option of pcretest has been documented for ages as being an old
1776        synonym of -m (show memory usage). I have changed it to mean "force study
1777        for every regex", that is, assume /S for every regex. This is similar to -i
1778        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1779        using it. It makes it easier to run collections of tests with and without
1780        study enabled, and thereby test pcre_study() more easily. All the standard
1781        tests are now run with and without -s (but some patterns can be marked as
1782        "never study" - see 20 below).
1783    
1784    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1785        restoration of the capturing data to the outer values was not happening
1786        correctly.
1787    
1788    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1789        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1790        pattern had matched an empty string, and so incorrectly returned a no
1791        match.
1792    
1793    17. There was optimizing code for the last branch of non-capturing parentheses,
1794        and also for the obeyed branch of a conditional subexpression, which used
1795        tail recursion to cut down on stack usage. Unfortunately, now that there is
1796        the possibility of (*THEN) occurring in these branches, tail recursion is
1797        no longer possible because the return has to be checked for (*THEN). These
1798        two optimizations have therefore been removed. [But see 8.20/11 above.]
1799    
1800    18. If a pattern containing \R was studied, it was assumed that \R always
1801        matched two bytes, thus causing the minimum subject length to be
1802        incorrectly computed because \R can also match just one byte.
1803    
1804    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1805        was incorrectly computed.
1806    
1807    20. If /S is present twice on a test pattern in pcretest input, it now
1808        *disables* studying, thereby overriding the use of -s on the command line
1809        (see 14 above). This is necessary for one or two tests to keep the output
1810        identical in both cases.
1811    
1812    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1813        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1814    
1815    22. When an atomic group that contained a capturing parenthesis was
1816        successfully matched, but the branch in which it appeared failed, the
1817        capturing was not being forgotten if a higher numbered group was later
1818        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1819        group 1 to "a", when in fact it should be unset. This applied to multi-
1820        branched capturing and non-capturing groups, repeated or not, and also to
1821        positive assertions (capturing in negative assertions does not happen
1822        in PCRE) and also to nested atomic groups.
1823    
1824    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1825        subject after a captured substring, to make it easier to tell which of a
1826        number of identical substrings has been captured.
1827    
1828    24. The way atomic groups are processed by pcre_exec() has been changed so that
1829        if they are repeated, backtracking one repetition now resets captured
1830        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1831        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1832        "aaa". Previously, it would have been "a". As part of this code
1833        refactoring, the way recursive calls are handled has also been changed.
1834    
1835    25. If an assertion condition captured any substrings, they were not passed
1836        back unless some other capturing happened later. For example, if
1837        (?(?=(a))a) was matched against "a", no capturing was returned.
1838    
1839    26. When studying a pattern that contained subroutine calls or assertions,
1840        the code for finding the minimum length of a possible match was handling
1841        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1842        group 1 called group 2 while simultaneously a separate group 2 called group
1843        1). A stack overflow occurred in this case. I have fixed this by limiting
1844        the recursion depth to 10.
1845    
1846    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1847        Fortmann. This supports explicit test numbers on the command line, and has
1848        argument validation and error reporting.
1849    
1850    28. An instance of \X with an unlimited repeat could fail if at any point the
1851        first character it looked at was a mark character.
1852    
1853    29. Some minor code refactoring concerning Unicode properties and scripts
1854        should reduce the stack requirement of match() slightly.
1855    
1856    30. Added the '=' option to pcretest to check the setting of unused capturing
1857        slots at the end of the pattern, which are documented as being -1, but are
1858        not included in the return count.
1859    
1860    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1861        compiled something random. Now it gives a compile-time error (as does
1862        Perl).
1863    
1864    32. A *MARK encountered during the processing of a positive assertion is now
1865        recorded and passed back (compatible with Perl).
1866    
1867    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1868        had alternative anchored branches, the search for a second match in a line
1869        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1870        matched the line "0102" twice. The same bug affected patterns that started
1871        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1872        twice.
1873    
1874    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1875        does, and because of capturing effects, quantifying parenthesized
1876        assertions may at times be useful. Quantifiers are now allowed for
1877        parenthesized assertions.
1878    
1879    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1880    
1881    36. \g was being checked for fancy things in a character class, when it should
1882        just be a literal "g".
1883    
1884    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1885        appearance of a nested POSIX class supersedes an apparent external class.
1886        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1887        unescaped square brackets may also appear as part of class names. For
1888        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1889        more like Perl. (But see 8.20/1 above.)
1890    
1891    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1892        was because it thought it was \N{name}, which is not supported).
1893    
1894    39. Add minix to OS list not supporting the -S option in pcretest.
1895    
1896    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1897        cannot analyze patterns in sufficient detail to catch mutual recursions
1898        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1899        subgroup is called recursively as a subpattern for a second time at the
1900        same position in the subject string. In previous releases this might have
1901        been caught by the recursion limit, or it might have run out of stack.
1902    
1903    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1904        happen only once. PCRE was, however incorrectly giving a compile time error
1905        "recursive call could loop indefinitely" because it cannot analyze the
1906        pattern in sufficient detail. The compile time test no longer happens when
1907        PCRE is compiling a conditional subpattern, but actual runaway loops are
1908        now caught at runtime (see 40 above).
1909    
1910    42. It seems that Perl allows any characters other than a closing parenthesis
1911        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1912        has been changed to be the same.
1913    
1914    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1915        as not to get warnings when autogen.sh is called. Also changed
1916        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1917    
1918    44. To help people who use pcregrep to scan files containing exceedingly long
1919        lines, the following changes have been made:
1920    
1921        (a) The default value of the buffer size parameter has been increased from
1922            8K to 20K. (The actual buffer used is three times this size.)
1923    
1924        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1925            PCRE is built.
1926    
1927        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1928            to be set at run time.
1929    
1930        (d) Numerical values in pcregrep options can be followed by K or M, for
1931            example --buffer-size=50K.
1932    
1933        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1934            given and the return code is set to 2.
1935    
1936    45. Add a pointer to the latest mark to the callout data block.
1937    
1938    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1939        partial match of an empty string instead of no match. This was specific to
1940        the use of ".".
1941    
1942    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1943        complete match instead of a partial match. This bug was dependent on both
1944        the PCRE_UTF8 and PCRE_DOTALL options being set.
1945    
1946    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1947        starting byte set, because \b was not being ignored.
1948    
1949    
1950    Version 8.12 15-Jan-2011
1951    ------------------------
1952    
1953    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1954        checks for such things as part of the documentation building process.
1955    
1956    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1957        --match-limit and --recursion-limit options (added for 8.11). In
1958        particular, this made one of the standard tests fail. (The integer value
1959        went into the wrong half of a long int.)
1960    
1961    3.  If the --colour option was given to pcregrep with -v (invert match), it
1962        did strange things, either producing crazy output, or crashing. It should,
1963        of course, ignore a request for colour when reporting lines that do not
1964        match.
1965    
1966    4.  Another pcregrep bug caused similar problems if --colour was specified with
1967        -M (multiline) and the pattern match finished with a line ending.
1968    
1969    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1970        matched in multiline mode, the following line was shown as part of the
1971        match. This seems wrong, so I have changed it.
1972    
1973    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1974        the check for further matches in the same line (so they could be coloured)
1975        to overrun the end of the current line. If another match was found, it was
1976        incorrectly shown (and then shown again when found in the next line).
1977    
1978    7.  If pcregrep was compiled under Windows, there was a reference to the
1979        function pcregrep_exit() before it was defined. I am assuming this was
1980        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1981        reported by a user. I've moved the definition above the reference.
1982    
1983    
1984    Version 8.11 10-Dec-2010
1985    ------------------------
1986    
1987    1.  (*THEN) was not working properly if there were untried alternatives prior
1988        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1989        backtracked to try for "b" instead of moving to the next alternative branch
1990        at the same level (in this case, to look for "c"). The Perl documentation
1991        is clear that when (*THEN) is backtracked onto, it goes to the "next
1992        alternative in the innermost enclosing group".
1993    
1994    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1995        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1996        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1997        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1998        (*THEN).
1999    
2000    3.  If \s appeared in a character class, it removed the VT character from
2001        the class, even if it had been included by some previous item, for example
2002        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2003        of \s, but is part of the POSIX "space" class.)
2004    
2005    4.  A partial match never returns an empty string (because you can always
2006        match an empty string at the end of the subject); however the checking for
2007        an empty string was starting at the "start of match" point. This has been
2008        changed to the "earliest inspected character" point, because the returned
2009        data for a partial match starts at this character. This means that, for
2010        example, /(?<=abc)def/ gives a partial match for the subject "abc"
2011        (previously it gave "no match").
2012    
2013    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2014        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2015        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2016        has an implication that the given string is incomplete (because a partial
2017        match is preferred over a full match). For this reason, these items now
2018        give a partial match in this situation. [Aside: previously, the one case
2019        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2020        match rather than a full match, which was wrong by the old rules, but is
2021        now correct.]
2022    
2023    6.  There was a bug in the handling of #-introduced comments, recognized when
2024        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2025        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2026        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2027        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2028        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2029        places in pcre_compile().
2030    
2031    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
2032        comments when looking ahead for named forward references to subpatterns,
2033        the only newline sequence it recognized was NL. It now handles newlines
2034        according to the set newline convention.
2035    
2036    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2037        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2038        cater for a lack of strerror(). These oversights have been fixed.
2039    
2040    9.  Added --match-limit and --recursion-limit to pcregrep.
2041    
2042    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2043    
2044    11. When the -o option was used, pcregrep was setting a return code of 1, even
2045        when matches were found, and --line-buffered was not being honoured.
2046    
2047    12. Added an optional parentheses number to the -o and --only-matching options
2048        of pcregrep.
2049    
2050    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2051        can match an empty string. The code to do it in pcretest and pcredemo
2052        needed fixing:
2053    
2054        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2055            only one byte after an empty string match just before CRLF (this case
2056            just got forgotten; "any" and "anycrlf" were OK).
2057    
2058        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2059            mode when an empty string match preceded an ASCII character followed by
2060            a non-ASCII character. (The code for advancing by one character rather
2061            than one byte was nonsense.)
2062    
2063        (c) The pcredemo.c sample program did not have any code at all to handle
2064            the cases when CRLF is a valid newline sequence.
2065    
2066    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2067        as a starting offset was within the subject string. There is now a new
2068        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2069        negative or greater than the length of the string. In order to test this,
2070        pcretest is extended to allow the setting of negative starting offsets.
2071    
2072    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2073        starting offset points to the beginning of a UTF-8 character was
2074        unnecessarily clumsy. I tidied it up.
2075    
2076    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2077        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2078    
2079    17. Nobody had reported that the --include_dir option, which was added in
2080        release 7.7 should have been called --include-dir (hyphen, not underscore)
2081        for compatibility with GNU grep. I have changed it to --include-dir, but
2082        left --include_dir as an undocumented synonym, and the same for
2083        --exclude-dir, though that is not available in GNU grep, at least as of
2084        release 2.5.4.
2085    
2086    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2087        characters from a string of bytes) have been redefined so as not to use
2088        loops, in order to improve performance in some environments. At the same
2089        time, I abstracted some of the common code into auxiliary macros to save
2090        repetition (this should not affect the compiled code).
2091    
2092    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2093        compile-time error is now given if \c is not followed by an ASCII
2094        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2095        different, and any byte value is allowed.)
2096    
2097    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2098        START_OPTIMIZE option, which is now allowed at compile time - but just
2099        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2100        to pcregrep and other applications that have no direct access to PCRE
2101        options. The new /Y option in pcretest sets this option when calling
2102        pcre_compile().
2103    
2104    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2105        back references. Groups containing recursive back references were forced to
2106        be atomic by that change, but in the case of named groups, the amount of
2107        memory required was incorrectly computed, leading to "Failed: internal
2108        error: code overflow". This has been fixed.
2109    
2110    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2111        pcretest.c, to avoid build problems in some Borland environments.
2112    
2113    
2114    Version 8.10 25-Jun-2010
2115    ------------------------
2116    
2117    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2118        THEN.
2119    
2120    2.  (*ACCEPT) was not working when inside an atomic group.
2121    
2122    3.  Inside a character class, \B is treated as a literal by default, but
2123        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2124        causes the error). The code is unchanged, but I tidied the documentation.
2125    
2126    4.  Inside a character class, PCRE always treated \R and \X as literals,
2127        whereas Perl faults them if its -w option is set. I have changed PCRE so
2128        that it faults them when PCRE_EXTRA is set.
2129    
2130    5.  Added support for \N, which always matches any character other than
2131        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2132    
2133    6.  When compiling pcregrep with newer versions of gcc which may have
2134        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2135        declared with attribute warn_unused_result" were given. Just casting the
2136        result to (void) does not stop the warnings; a more elaborate fudge is
2137        needed. I've used a macro to implement this.
2138    
2139    7.  Minor change to pcretest.c to avoid a compiler warning.
2140    
2141    8.  Added four artifical Unicode properties to help with an option to make
2142        \s etc use properties (see next item). The new properties are: Xan
2143        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2144    
2145    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2146        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2147        this option. Modified pcretest to add /W to test this facility. Added
2148        REG_UCP to make it available via the POSIX interface.
2149    
2150    10. Added --line-buffered to pcregrep.
2151    
2152    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2153        studied, and the match started with a letter with a code point greater than
2154        127 whose first byte was different to the first byte of the other case of
2155        the letter, the other case of this starting letter was not recognized
2156        (#976).
2157    
2158    12. If a pattern that was studied started with a repeated Unicode property
2159        test, for example, \p{Nd}+, there was the theoretical possibility of
2160        setting up an incorrect bitmap of starting bytes, but fortunately it could
2161        not have actually happened in practice until change 8 above was made (it
2162        added property types that matched character-matching opcodes).
2163    
2164    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2165        possible starting bytes for non-anchored patterns.
2166    
2167    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2168        \R, and also a number of cases that involve Unicode properties, both
2169        explicit and implicit when PCRE_UCP is set.
2170    
2171    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2172        input, it could crash or give wrong results if characters with values
2173        greater than 0xc0 were present in the subject string. (Detail: it assumed
2174        UTF-8 input when processing these items.)
2175    
2176    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2177        size_t is 64-bit (#991).
2178    
2179    17. Added a check for running out of memory when PCRE is compiled with
2180        --disable-stack-for-recursion (#990).
2181    
2182    18. If the last data line in a file for pcretest does not have a newline on
2183        the end, a newline was missing in the output.
2184    
2185    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2186        less than 128) in its various bitmaps. However, there is a facility for
2187        generating tables according to the current locale when PCRE is compiled. It
2188        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2189        characters, are recognized by isspace() and therefore were getting set in
2190        these tables, and indeed these tables seem to approximate to ISO 8859. This
2191        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2192        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2193        which of course cannot start UTF-8 characters. I have changed the code so
2194        that only real ASCII characters (less than 128) and the correct starting
2195        bytes for UTF-8 encodings are set for characters greater than 127 when in
2196        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2197        altogether.)
2198    
2199    20. Added the /T option to pcretest so as to be able to run tests with non-
2200        standard character tables, thus making it possible to include the tests
2201        used for 19 above in the standard set of tests.
2202    
2203    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2204        reference to a subpattern the other side of a comment that contains an
2205        opening parenthesis caused either an internal compiling error, or a
2206        reference to the wrong subpattern.
2207    
2208    
2209    Version 8.02 19-Mar-2010
2210    ------------------------
2211    
2212    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2213    
2214    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2215        configured.
2216    
2217    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2218        original author of that file, following a query about its status.
2219    
2220    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2221        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2222    
2223    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2224        quantifier applied to a forward-referencing subroutine call, could compile
2225        incorrect code or give the error "internal error: previously-checked
2226        referenced subpattern not found".
2227    
2228    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2229        variables to point to external functions. For these systems, therefore,
2230        pcre_malloc etc. are now initialized to local functions that call the
2231        relevant global functions.
2232    
2233    7.  There were two entries missing in the vectors called coptable and poptable
2234        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2235        I've fixed the data, and added a kludgy way of testing at compile time that
2236        the lengths are correct (equal to the number of opcodes).
2237    
2238    8.  Following on from 7, I added a similar kludge to check the length of the
2239        eint vector in pcreposix.c.
2240    
2241    9.  Error texts for pcre_compile() are held as one long string to avoid too
2242        much relocation at load time. To find a text, the string is searched,
2243        counting zeros. There was no check for running off the end of the string,
2244        which could happen if a new error number was added without updating the
2245        string.
2246    
2247    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2248    
2249    11. \K was not working if it appeared in an atomic group or in a group that
2250        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2251        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2252        the assertion is positive, but not if it is negative.
2253    
2254    12. Change 11 fortuitously reduced the size of the stack frame used in the
2255        "match()" function of pcre_exec.c by one pointer. Forthcoming
2256        implementation of support for (*MARK) will need an extra pointer on the
2257        stack; I have reserved it now, so that the stack frame size does not
2258        decrease.
2259    
2260    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2261        item in branch that calls a recursion is a subroutine call - as in the
2262        second branch in the above example - was incorrectly given the compile-
2263        time error "recursive call could loop indefinitely" because pcre_compile()
2264        was not correctly checking the subroutine for matching a non-empty string.
2265    
2266    14. The checks for overrunning compiling workspace could trigger after an
2267        overrun had occurred. This is a "should never occur" error, but it can be
2268        triggered by pathological patterns such as hundreds of nested parentheses.
2269        The checks now trigger 100 bytes before the end of the workspace.
2270    
2271    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2272    
2273    
2274    Version 8.01 19-Jan-2010
2275    ------------------------
2276    
2277    1.  If a pattern contained a conditional subpattern with only one branch (in
2278        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2279        computed the wrong minimum data length (which is of course zero for such
2280        subpatterns). This could cause incorrect "no match" results.
2281    
2282    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2283        the pattern is reset in the first branch, pcre_compile() failed with
2284        "internal error: code overflow at offset...". This happened only when
2285        the reset was to the original external option setting. (An optimization
2286        abstracts leading options settings into an external setting, which was the
2287        cause of this.)
2288    
2289    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2290        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2291        assertion pattern did not match (meaning that the assertion was true), it
2292        was incorrectly treated as false if the SKIP had been reached during the
2293        matching. This also applied to assertions used as conditions.
2294    
2295    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2296        assertion subpattern, including such a pattern used as a condition,
2297        unpredictable results occurred, instead of the error return
2298        PCRE_ERROR_DFA_UITEM.
2299    
2300    5.  The C++ GlobalReplace function was not working like Perl for the special
2301        situation when an empty string is matched. It now does the fancy magic
2302        stuff that is necessary.
2303    
2304    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2305        removed. (These were left over from very, very early versions of PCRE.)
2306    
2307    7.  Some cosmetic changes to the code to make life easier when compiling it
2308        as part of something else:
2309    
2310        (a) Change DEBUG to PCRE_DEBUG.
2311    
2312        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2313            called "current" as "current_branch", to prevent a collision with the
2314            Linux macro when compiled as a kernel module.
2315    
2316        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2317            prevent a collision with the Linux macro when compiled as a kernel
2318            module.
2319    
2320    8.  In pcre_compile() there are some checks for integer overflows that used to
2321        cast potentially large values to (double). This has been changed to that
2322        when building, a check for int64_t is made, and if it is found, it is used
2323        instead, thus avoiding the use of floating point arithmetic. (There is no
2324        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2325        double.
2326    
2327    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2328        2005 (difference between two addresses compared to an unsigned value).
2329    
2330    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2331        custom one, because of the following reported problem in Windows:
2332    
2333          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2334              under Win32.
2335          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2336              therefore missing the function definition.
2337          - The compiler thus generates a "C" signature for the test function.
2338          - The linker fails to find the "C" function.
2339          - PCRE fails to configure if asked to do so against libbz2.
2340    
2341    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2342        messages were output:
2343    
2344          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2345          rerunning libtoolize, to keep the correct libtool macros in-tree.
2346          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2347    
2348        I have done both of these things.
2349    
2350    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2351        most of the time, it *can* run out if it is given a pattern that contains a
2352        runaway infinite recursion. I updated the discussion in the pcrestack man
2353        page.
2354    
2355    13. Now that we have gone to the x.xx style of version numbers, the minor
2356        version may start with zero. Using 08 or 09 is a bad idea because users
2357        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2358        interpreted as invalid octal numbers. I've updated the previous comment in
2359        configure.ac, and also added a check that gives an error if 08 or 09 are
2360        used.
2361    
2362    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2363        causing partial matching to fail when the end of the subject matched \W
2364        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2365    
2366    15. There were some discrepancies between the declarations in pcre_internal.h
2367        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2368        their definitions. The declarations used "const uschar *" and the
2369        definitions used USPTR. Even though USPTR is normally defined as "const
2370        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2371        reported that: "This difference in casting confuses some C++ compilers, for
2372        example, SunCC recognizes above declarations as different functions and
2373        generates broken code for hbpcre." I have changed the declarations to use
2374        USPTR.
2375    
2376    16. GNU libtool is named differently on some systems. The autogen.sh script now
2377        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2378        (FreeBSD).
2379    
2380    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2381        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2382        comment: "Figure out how to create a longlong from a string: strtoll and
2383        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2384        instance, but it only takes 2 args instead of 3!"
2385    
2386    18. A subtle bug concerned with back references has been fixed by a change of
2387        specification, with a corresponding code fix. A pattern such as
2388        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2389        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2390        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2391        same bug. Such groups have to be quantified to be useful, or contained
2392        inside another quantified group. (If there's no repetition, the reference
2393        can never match.) The problem arises because, having left the group and
2394        moved on to the rest of the pattern, a later failure that backtracks into
2395        the group uses the captured value from the final iteration of the group
2396        rather than the correct earlier one. I have fixed this in PCRE by forcing
2397        any group that contains a reference to itself to be an atomic group; that
2398        is, there cannot be any backtracking into it once it has completed. This is
2399        similar to recursive and subroutine calls.
2400    
2401    
2402    Version 8.00 19-Oct-09
2403    ----------------------
2404    
2405    1.  The table for translating pcre_compile() error codes into POSIX error codes
2406        was out-of-date, and there was no check on the pcre_compile() error code
2407        being within the table. This could lead to an OK return being given in
2408        error.
2409    
2410    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2411        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2412        in a Windows environment.
2413    
2414    3.  The pcregrep --count option prints the count for each file even when it is
2415        zero, as does GNU grep. However, pcregrep was also printing all files when
2416        --files-with-matches was added. Now, when both options are given, it prints
2417        counts only for those files that have at least one match. (GNU grep just
2418        prints the file name in this circumstance, but including the count seems
2419        more useful - otherwise, why use --count?) Also ensured that the
2420        combination -clh just lists non-zero counts, with no names.
2421    
2422    4.  The long form of the pcregrep -F option was incorrectly implemented as
2423        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2424        but it seems right to fix it, and I didn't think it was worth preserving
2425        the old behaviour.
2426    
2427    5.  The command line items --regex=pattern and --regexp=pattern were not
2428        recognized by pcregrep, which required --regex pattern or --regexp pattern
2429        (with a space rather than an '='). The man page documented the '=' forms,
2430        which are compatible with GNU grep; these now work.
2431    
2432    6.  No libpcreposix.pc file was created for pkg-config; there was just
2433        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2434    
2435    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2436        when UCP support is not needed, by modifying the Python script that
2437        generates it from Unicode data files. This should not matter if the module
2438        is correctly used as a library, but I received one complaint about 50K of
2439        unwanted data. My guess is that the person linked everything into his
2440        program rather than using a library. Anyway, it does no harm.
2441    
2442    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2443        was a minimum greater than 1 for a wide character in a possessive
2444        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2445        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2446        character. Chaos in the form of incorrect output or a compiling loop could
2447        result.
2448    
2449    9.  The restrictions on what a pattern can contain when partial matching is
2450        requested for pcre_exec() have been removed. All patterns can now be
2451        partially matched by this function. In addition, if there are at least two
2452        slots in the offset vector, the offset of the earliest inspected character
2453        for the match and the offset of the end of the subject are set in them when
2454        PCRE_ERROR_PARTIAL is returned.
2455    
2456    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2457        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2458        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2459        and may be more useful for multi-segment matching.
2460    
2461    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2462        used to be given if ever the end of the subject was reached; now it is
2463        given only if matching could not proceed because another character was
2464        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2465        string "Z", which now yields "no match" instead of "partial match". In the
2466        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2467        final character ended with (*FAIL).
2468    
2469    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2470        if the pattern had a "must contain" character that was already found in the
2471        earlier partial match, unless partial matching was again requested. For
2472        example, with the pattern /dog.(body)?/, the "must contain" character is
2473        "g". If the first part-match was for the string "dog", restarting with
2474        "sbody" failed. This bug has been fixed.
2475    
2476    13. The string returned by pcre_dfa_exec() after a partial match has been
2477        changed so that it starts at the first inspected character rather than the
2478        first character of the match. This makes a difference only if the pattern
2479        starts with a lookbehind assertion or \b or \B (\K is not supported by
2480        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2481        matching functions compatible, and I think it's the right thing to do.
2482    
2483    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2484        so that the demonstration program is easily available in environments where
2485        PCRE has not been installed from source.
2486    
2487    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2488        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2489        library.
2490    
2491    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2492        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2493        is not the first non-POSIX option to be added. Clearly some people find
2494        these options useful.
2495    
2496    17. If a caller to the POSIX matching function regexec() passes a non-zero
2497        value for nmatch with a NULL value for pmatch, the value of
2498        nmatch is forced to zero.
2499    
2500    18. RunGrepTest did not have a test for the availability of the -u option of
2501        the diff command, as RunTest does. It now checks in the same way as
2502        RunTest, and also checks for the -b option.
2503    
2504    19. If an odd number of negated classes containing just a single character
2505        interposed, within parentheses, between a forward reference to a named
2506        subpattern and the definition of the subpattern, compilation crashed with
2507        an internal error, complaining that it could not find the referenced
2508        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2509        [The bug was that it was starting one character too far in when skipping
2510        over the character class, thus treating the ] as data rather than
2511        terminating the class. This meant it could skip too much.]
2512    
2513    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2514        /g option in pcretest when the pattern contains \K, which makes it possible
2515        to have an empty string match not at the start, even when the pattern is
2516        anchored. Updated pcretest and pcredemo to use this option.
2517    
2518    21. If the maximum number of capturing subpatterns in a recursion was greater
2519        than the maximum at the outer level, the higher number was returned, but
2520        with unset values at the outer level. The correct (outer level) value is
2521        now given.
2522    
2523    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2524        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2525        make it do so. The string so far is captured, making this feature
2526        compatible with Perl.
2527    
2528    23. The tests have been re-organized, adding tests 11 and 12, to make it
2529        possible to check the Perl 5.10 features against Perl 5.10.
2530    
2531    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2532        pattern matches a fixed length string. PCRE did not allow this; now it
2533        does. Neither allows recursion.
2534    
2535    25. I finally figured out how to implement a request to provide the minimum
2536        length of subject string that was needed in order to match a given pattern.
2537        (It was back references and recursion that I had previously got hung up
2538        on.) This code has now been added to pcre_study(); it finds a lower bound
2539        to the length of subject needed. It is not necessarily the greatest lower
2540        bound, but using it to avoid searching strings that are too short does give
2541        some useful speed-ups. The value is available to calling programs via
2542        pcre_fullinfo().
2543    
2544    26. While implementing 25, I discovered to my embarrassment that pcretest had
2545        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2546        study optimizations had never been tested with that matching function.
2547        Oops. What is worse, even when it was passed study data, there was a bug in
2548        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2549        were also very few tests of studied patterns with pcre_dfa_exec().
2550    
2551    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2552        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2553        on the other side of the coin, they are no longer allowed to have different
2554        names, because these cannot be distinguished in PCRE, and this has caused
2555        confusion. (This is a difference from Perl.)
2556    
2557    28. When duplicate subpattern names are present (necessarily with different
2558        numbers, as required by 27 above), and a test is made by name in a
2559        conditional pattern, either for a subpattern having been matched, or for
2560        recursion in such a pattern, all the associated numbered subpatterns are
2561        tested, and the overall condition is true if the condition is true for any
2562        one of them. This is the way Perl works, and is also more like the way
2563        testing by number works.
2564    
2565    
2566    Version 7.9 11-Apr-09
2567    ---------------------
2568    
2569    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2570        (pcretest), all targets were linked against these libraries. This included
2571        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2572        libraries. This caused unwanted dependencies to be created. This problem
2573        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2574        pcretest is linked with readline.
2575    
2576    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2577        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2578        moved outside it again, because FALSE and TRUE are already defined in AIX,
2579        but BOOL is not.
2580    
2581    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2582        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2583    
2584    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2585        hyphens) following file names and line numbers when outputting matching
2586        lines. This is not true; no spaces are inserted. I have also clarified the
2587        wording for the --colour (or --color) option.
2588    
2589    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2590        was not coloured; this is different to GNU grep, so I have changed it to be
2591        the same.
2592    
2593    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2594        each matching line was coloured. Now it goes on to look for further matches
2595        of any of the test patterns, which is the same behaviour as GNU grep.
2596    
2597    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2598        doesn't make sense to accept an empty string match in pcregrep, so I have
2599        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2600        seems to be how GNU grep behaves. [But see later change 40 for release
2601        8.33.]
2602    
2603    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2604        start or after a newline", because the conditional assertion was not being
2605        correctly handled. The rule now is that both the assertion and what follows
2606        in the first alternative must satisfy the test.
2607    
2608    9.  If auto-callout was enabled in a pattern with a conditional group whose
2609        condition was an assertion, PCRE could crash during matching, both with
2610        pcre_exec() and pcre_dfa_exec().
2611    
2612    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2613        used for matching.
2614    
2615    11. Unicode property support in character classes was not working for
2616        characters (bytes) greater than 127 when not in UTF-8 mode.
2617    
2618    12. Added the -M command line option to pcretest.
2619    
2620    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2621    
2622    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2623    
2624    16. Added comments and documentation about mis-use of no_arg in the C++
2625        wrapper.
2626    
2627    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2628        from Martin Jerabek that uses macro names for all relevant character and
2629        string constants.
2630    
2631    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2632        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2633        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2634        these, but not everybody uses configure.
2635    
2636    19. A conditional group that had only one branch was not being correctly
2637        recognized as an item that could match an empty string. This meant that an
2638        enclosing group might also not be so recognized, causing infinite looping
2639        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2640        with the subject "ab", where knowledge that the repeated group can match
2641        nothing is needed in order to break the loop.
2642    
2643    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2644        exec(), but without supplying a callout function, matching went wrong.
2645    
2646    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2647        leak if the size of the offset vector was greater than 30. When the vector
2648        is smaller, the saved offsets during recursion go onto a local stack
2649        vector, but for larger vectors malloc() is used. It was failing to free
2650        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2651        error, in fact).
2652    
2653    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2654        heapframe that is used only when UTF-8 support is enabled. This caused no
2655        problem, but was untidy.
2656    
2657    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2658        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2659        included within another project.
2660    
2661    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2662        slightly modified by me:
2663    
2664          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2665              not building pcregrep.
2666    
2667          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2668              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2669    
2670    25. Forward references, both numeric and by name, in patterns that made use of
2671        duplicate group numbers, could behave incorrectly or give incorrect errors,
2672        because when scanning forward to find the reference group, PCRE was not
2673        taking into account the duplicate group numbers. A pattern such as
2674        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2675    
2676    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2677        the feature of a custom pointer more persuasive (as requested by a user).
2678    
2679    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2680        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2681        pre-defined.
2682    
2683    28. Added support for (*UTF8) at the start of a pattern.
2684    
2685    29. Arrange for flags added by the "release type" setting in CMake to be shown
2686        in the configuration summary.
2687    
2688    
2689    Version 7.8 05-Sep-08
2690    ---------------------
2691    
2692    1.  Replaced UCP searching code with optimized version as implemented for Ad
2693        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2694        stage table and inline lookup instead of a function, giving speed ups of 2
2695        to 5 times on some simple patterns that I tested. Permission was given to
2696        distribute the MultiStage2.py script that generates the tables (it's not in
2697        the tarball, but is in the Subversion repository).
2698    
2699    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2700        scripts.
2701    
2702    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2703        a group with a zero qualifier. The result of the study could be incorrect,
2704        or the function might crash, depending on the pattern.
2705    
2706    4.  Caseless matching was not working for non-ASCII characters in back
2707        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2708        It now works when Unicode Property Support is available.
2709    
2710    5.  In pcretest, an escape such as \x{de} in the data was always generating
2711        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2712        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2713        truncation.
2714    
2715    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2716    
2717    7.  Added two (int) casts to pcregrep when printing the difference of two
2718        pointers, in case they are 64-bit values.
2719    
2720    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2721        test 2 if it fails.
2722    
2723    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2724        and a #define of that name to empty if it is not externally set. This is to
2725        allow users of MSVC to set it if necessary.
2726    
2727    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2728        the convenience functions in the pcre_get.c source file.
2729    
2730    11. An option change at the start of a pattern that had top-level alternatives
2731        could cause overwriting and/or a crash. This command provoked a crash in
2732        some environments:
2733    
2734          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2735    
2736        This potential security problem was recorded as CVE-2008-2371.
2737    
2738    12. For a pattern where the match had to start at the beginning or immediately
2739        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2740        pcre_dfa_exec() could read past the end of the passed subject if there was
2741        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2742        pcretest so that it places the subject at the end of its malloc-ed buffer.
2743    
2744    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2745        exec() might read past the end of the data buffer in UTF-8 mode.
2746    
2747    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2748        the data contained the byte 0x85 as part of a UTF-8 character within its
2749        first line. This applied both to normal and DFA matching.
2750    
2751    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2752        /^[^d]*?$/8 failed to match "abc".
2753    
2754    16. Added a missing copyright notice to pcrecpp_internal.h.
2755    
2756    17. Make it more clear in the documentation that values returned from
2757        pcre_exec() in ovector are byte offsets, not character counts.
2758    
2759    18. Tidied a few places to stop certain compilers from issuing warnings.
2760    
2761    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2762        supplied by Stefan Weber. I made a further small update for 7.8 because
2763        there is a change of source arrangements: the pcre_searchfuncs.c module is
2764        replaced by pcre_ucd.c.
2765    
2766    
2767    Version 7.7 07-May-08
2768    ---------------------
2769    
2770    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2771        a string to a long long, pretend we don't even have a long long." This is
2772        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2773    
2774    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2775        pre-7.6 versions, which defined a global no_arg variable instead of putting
2776        it in the RE class. (See also #8 below.)
2777    
2778    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2779        Lopes.
2780    
2781    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2782    
2783        (1) The include/exclude patterns were being applied to the whole pathnames
2784            of files, instead of just to the final components.
2785    
2786        (2) If there was more than one level of directory, the subdirectories were
2787            skipped unless they satisfied the include/exclude conditions. This is
2788            inconsistent with GNU grep (and could even be seen as contrary to the
2789            pcregrep specification - which I improved to make it absolutely clear).
2790            The action now is always to scan all levels of directory, and just
2791            apply the include/exclude patterns to regular files.
2792    
2793    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2794        --exclude_dir in the tests to avoid scanning .svn directories.
2795    
2796    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2797        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2798        doesn't support NULs in patterns.
2799    
2800    7.  Added some missing "const"s to declarations of static tables in
2801        pcre_compile.c and pcre_dfa_exec.c.
2802    
2803    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2804        caused by fix #2  above. (Subsequently also a second patch to fix the
2805        first patch. And a third patch - this was a messy problem.)
2806    
2807    9.  Applied Craig's patch to remove the use of push_back().
2808    
2809    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2810        matching function regexec().
2811    
2812    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2813        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2814        references. PCRE supports relative numbers with this syntax (I don't think
2815        Oniguruma does).
2816    
2817    12. Previously, a group with a zero repeat such as (...){0} was completely
2818        omitted from the compiled regex. However, this means that if the group
2819        was called as a subroutine from elsewhere in the pattern, things went wrong
2820        (an internal error was given). Such groups are now left in the compiled
2821        pattern, with a new opcode that causes them to be skipped at execution
2822        time.
2823    
2824    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2825        to the way PCRE behaves:
2826    
2827        (a) A lone ] character is dis-allowed (Perl treats it as data).
2828    
2829        (b) A back reference to an unmatched subpattern matches an empty string
2830            (Perl fails the current match path).
2831    
2832        (c) A data ] in a character class must be notated as \] because if the
2833            first data character in a class is ], it defines an empty class. (In
2834            Perl it is not possible to have an empty class.) The empty class []
2835            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2836            The negative empty class [^] matches any one character, independently
2837            of the DOTALL setting.
2838    
2839    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2840        non-existent subpattern following a character class starting with ']' and
2841        containing () gave an internal compiling error instead of "reference to
2842        non-existent subpattern". Fortunately, when the pattern did exist, the
2843        compiled code was correct. (When scanning forwards to check for the
2844        existence of the subpattern, it was treating the data ']' as terminating
2845        the class, so got the count wrong. When actually compiling, the reference
2846        was subsequently set up correctly.)
2847    
2848    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2849        it was being rejected as not supported by pcre_dfa_exec(), even though
2850        other assertions are supported. I have made pcre_dfa_exec() support
2851        (*FAIL).
2852    
2853    16. The implementation of 13c above involved the invention of a new opcode,
2854        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2855        cannot be changed at match time, I realized I could make a small
2856        improvement to matching performance by compiling OP_ALLANY instead of
2857        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2858        on the OP_ANY path.
2859    
2860    17. Compiling pcretest on Windows with readline support failed without the
2861        following two fixes: (1) Make the unistd.h include conditional on
2862        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2863    
2864    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2865        ncurses library to be included for pcretest when ReadLine support is
2866        requested, but also to allow for it to be overridden. This patch came from
2867        Daniel Bergström.
2868    
2869    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2870        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2871        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2872        spotting this.
2873    
2874    
2875    Version 7.6 28-Jan-08
2876    ---------------------
2877    
2878    1.  A character class containing a very large number of characters with
2879        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2880        overflow.
2881    
2882    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2883        HAVE_LONG_LONG is not defined.
2884    
2885    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2886        bring them up to date and include new features. This patch includes:
2887    
2888        - Fixed PH's badly added libz and libbz2 support.
2889        - Fixed a problem with static linking.
2890        - Added pcredemo. [But later removed - see 7 below.]
2891        - Fixed dftables problem and added an option.
2892        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2893            HAVE_LONG_LONG.
2894        - Added readline support for pcretest.
2895        - Added an listing of the option settings after cmake has run.
2896    
2897    4.  A user submitted a patch to Makefile that makes it easy to create
2898        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2899        Makefile.am that cause it to include this special target, without
2900        affecting anything else. Note that the same mingw target plus all
2901        the other distribution libraries and programs are now supported
2902        when configuring with CMake (see 6 below) instead of with
2903        Configure/Make.
2904    
2905    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2906        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2907        exported in the Windows port". It has not yet been confirmed that the patch
2908        solves the problem, but it does no harm.
2909    
2910    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2911        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2912        with CMake, and also correct the comment about stack recursion.
2913    
2914    7.  Remove the automatic building of pcredemo from the ./configure system and
2915        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2916        of a program that users should build themselves after PCRE is installed, so
2917        building it automatically is not really right. What is more, it gave
2918        trouble in some build environments.
2919    
2920    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2921    
2922    
2923    Version 7.5 10-Jan-08
2924    ---------------------
2925    
2926    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2927        values in parens when parsing an RE using the C++ wrapper."
2928    
2929    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2930        Characters greater than 255 were excluded from the class instead of being
2931        included.
2932    
2933    3.  The same bug as (2) above applied to negated POSIX classes such as
2934        [:^space:].
2935    
2936    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2937        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2938        I have changed it.
2939    
2940    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2941        first named subpattern) and a construct such as (?&a) would reference the
2942        first named subpattern whose name started with "a" (in other words, the
2943        length check was missing). Both these problems are fixed. "Subpattern name
2944        expected" is now given for (?&) (a zero-length name), and this patch also
2945        makes it give the same error for \k'' (previously it complained that that
2946        was a reference to a non-existent subpattern).
2947    
2948    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2949        this is right because (?- can be followed by option settings as well as by
2950        digits. I have, however, made the messages clearer.
2951    
2952    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2953        than the number used in the conditional) now cause a compile-time error.
2954        This is actually not compatible with Perl, which accepts such patterns, but
2955        treats the conditional as always being FALSE (as PCRE used to), but it
2956        seems to me that giving a diagnostic is better.
2957    
2958    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2959        and messages.
2960    
2961    9.  Fix two occurrences of "backslash" in comments that should have been
2962        "backspace".
2963    
2964    10. Remove two redundant lines of code that can never be obeyed (their function
2965        was moved elsewhere).
2966    
2967    11. The program that makes PCRE's Unicode character property table had a bug
2968        which caused it to generate incorrect table entries for sequences of
2969        characters that have the same character type, but are in different scripts.
2970        It amalgamated them into a single range, with the script of the first of
2971        them. In other words, some characters were in the wrong script. There were
2972        thirteen such cases, affecting characters in the following ranges:
2973    
2974          U+002b0 - U+002c1
2975          U+0060c - U+0060d
2976          U+0061e - U+00612
2977          U+0064b - U+0065e
2978          U+0074d - U+0076d
2979          U+01800 - U+01805
2980          U+01d00 - U+01d77
2981          U+01d9b - U+01dbf
2982          U+0200b - U+0200f
2983          U+030fc - U+030fe
2984          U+03260 - U+0327f
2985          U+0fb46 - U+0fbb1
2986          U+10450 - U+1049d
2987    
2988    12. The -o option (show only the matching part of a line) for pcregrep was not
2989        compatible with GNU grep in that, if there was more than one match in a
2990        line, it showed only the first of them. It now behaves in the same way as
2991        GNU grep.
2992    
2993    13. If the -o and -v options were combined for pcregrep, it printed a blank
2994        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2995        does the same. The return code can be used to tell if there were any
2996        non-matching lines.
2997    
2998    14. Added --file-offsets and --line-offsets to pcregrep.
2999    
3000    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
3001        infinitely looping recursion. The bug was that positive lookaheads were not
3002        being skipped when checking for a possible empty match (negative lookaheads
3003        and both kinds of lookbehind were skipped).
3004    
3005    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
3006        inclusion of <windows.h> to before rather than after the definition of
3007        INVALID_FILE_ATTRIBUTES (patch from David Byron).
3008    
3009    17. Specifying a possessive quantifier with a specific limit for a Unicode
3010        character property caused pcre_compile() to compile bad code, which led at
3011        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
3012        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
3013        caused the error; without that there was no problem.
3014    
3015    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
3016    
3017    19. Added --enable-pcretest-libreadline.
3018    
3019    20. In pcrecpp.cc, the variable 'count' was incremented twice in
3020        RE::GlobalReplace(). As a result, the number of replacements returned was
3021        double what it should be. I removed one of the increments, but Craig sent a
3022        later patch that removed the other one (the right fix) and added unit tests
3023        that check the return values (which was not done before).
3024    
3025    21. Several CMake things:
3026    
3027        (1) Arranged that, when cmake is used on Unix, the libraries end up with
3028            the names libpcre and libpcreposix, not just pcre and pcreposix.
3029    
3030        (2) The above change means that pcretest and pcregrep are now correctly
3031            linked with the newly-built libraries, not previously installed ones.
3032    
3033        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
3034    
3035    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
3036        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
3037        UTF-8 newline character). The key issue is that the pattern starts .*;
3038        this means that the match must be either at the beginning, or after a
3039        newline. The bug was in the code for advancing after a failed match and
3040        checking that the new position followed a newline. It was not taking
3041        account of UTF-8 characters correctly.
3042    
3043    23. PCRE was behaving differently from Perl in the way it recognized POSIX
3044        character classes. PCRE was not treating the sequence [:...:] as a
3045        character class unless the ... were all letters. Perl, however, seems to
3046        allow any characters between [: and :], though of course it rejects as
3047        unknown any "names" that contain non-letters, because all the known class
3048        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
3049        for example, whereas PCRE did not - it did not recognize a POSIX character
3050        class. This seemed a bit dangerous, so the code has been changed to be
3051        closer to Perl. The behaviour is not identical to Perl, because PCRE will
3052        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3053        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3054        Perl does, and where it didn't before.
3055    
3056    24. Rewrite so as to remove the single use of %n from pcregrep because in some
3057        Windows environments %n is disabled by default.
3058    
3059    
3060    Version 7.4 21-Sep-07
3061    ---------------------
3062    
3063    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
3064        means that a class such as [\s] counted as "explicit reference to CR or
3065        LF". That isn't really right - the whole point of the change was to try to
3066        help when there was an actual mention of one of the two characters. So now
3067        the change happens only if \r or \n (or a literal CR or LF) character is
3068        encountered.
3069    
3070    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
3071        of both had grown to the point where there were only 3 bits left.
3072        Fortunately, there was spare space in the data structure, and so I have
3073        moved the internal flags into a new 16-bit field to free up more option
3074        bits.
3075    
3076    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3077        but did not set the internal JCHANGED flag - either of these is enough to
3078        control the way the "get" function works - but the PCRE_INFO_JCHANGED
3079        facility is supposed to tell if (?J) was ever used, so now (?J) at the
3080        start sets both bits.
3081    
3082    4.  Added options (at build time, compile time, exec time) to change \R from
3083        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3084    
3085    5.  doc/pcresyntax.html was missing from the distribution.
3086    
3087    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3088        compatibility, even though it is no longer used.
3089    
3090    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3091        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3092        windows.h file is present (where different names are used). [This was
3093        reversed later after testing - see 16 below.]
3094    
3095    8.  Changed all #include <config.h> to #include "config.h". There were also
3096        some further <pcre.h> cases that I changed to "pcre.h".
3097    
3098    9.  When pcregrep was used with the --colour option, it missed the line ending
3099        sequence off the lines that it output.
3100    
3101    10. It was pointed out to me that arrays of string pointers cause lots of
3102        relocations when a shared library is dynamically loaded. A technique of
3103        using a single long string with a table of offsets can drastically reduce
3104        these. I have refactored PCRE in four places to do this. The result is
3105        dramatic:
3106    
3107          Originally:                          290
3108          After changing UCP table:            187
3109          After changing error message table:   43
3110          After changing table of "verbs"       36
3111          After changing table of Posix names   22
3112    
3113        Thanks to the folks working on Gregex for glib for this insight.
3114    
3115    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3116        unicode-properties was also set.
3117    
3118    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3119    
3120    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3121        checked only for CRLF.
3122    
3123    14. Added casts to pcretest.c to avoid compiler warnings.
3124    
3125    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3126    
3127    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3128        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3129        entirely. This removes changes made in 7 above.
3130    
3131    17. The CMake files have been updated, and there is now more information about
3132        building with CMake in the NON-UNIX-USE document.
3133    
3134    
3135    Version 7.3 28-Aug-07
3136    ---------------------
3137    
3138     1. In the rejigging of the build system that eventually resulted in 7.1, the
3139        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3140        brackets there is not right, since it causes compilers to look for an
3141        installed pcre.h, not the version that is in the source that is being
3142        compiled (which of course may be different). I have changed it back to:
3143    
3144          #include "pcre.h"
3145    
3146        I have a vague recollection that the change was concerned with compiling in
3147        different directories, but in the new build system, that is taken care of
3148        by the VPATH setting the Makefile.
3149    
3150     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3151        when the subject happened to end in the byte 0x85 (e.g. if the last
3152        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3153        characters but of course it shouldn't be taken as a newline when it is part
3154        of another character. The bug was that, for an unlimited repeat of . in
3155        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3156        characters when looking for a newline.
3157    
3158     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3159    
3160     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3161        in debug output.
3162    
3163     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3164        long printing in the pcrecpp unittest when running under MinGW.
3165    
3166     6. ESC_K was left out of the EBCDIC table.
3167    
3168     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3169        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3170        limit also applies to "virtual nesting" when a pattern is recursive, and in
3171        this case 1000 isn't so big. I have been able to remove this limit at the
3172        expense of backing off one optimization in certain circumstances. Normally,
3173        when pcre_exec() would call its internal match() function recursively and
3174        immediately return the result unconditionally, it uses a "tail recursion"
3175        feature to save stack. However, when a subpattern that can match an empty
3176        string has an unlimited repetition quantifier, it no longer makes this
3177        optimization. That gives it a stack frame in which to save the data for
3178        checking that an empty string has been matched. Previously this was taken
3179        from the 1000-entry workspace that had been reserved. So now there is no
3180        explicit limit, but more stack is used.
3181    
3182     8. Applied Daniel's patches to solve problems with the import/export magic
3183        syntax that is required for Windows, and which was going wrong for the
3184        pcreposix and pcrecpp parts of the library. These were overlooked when this
3185        problem was solved for the main library.
3186    
3187     9. There were some crude static tests to avoid integer overflow when computing
3188        the size of patterns that contain repeated groups with explicit upper
3189        limits. As the maximum quantifier is 65535, the maximum group length was
3190        set at 30,000 so that the product of these two numbers did not overflow a
3191        32-bit integer. However, it turns out that people want to use groups that
3192        are longer than 30,000 bytes (though not repeat them that many times).
3193        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3194        made it possible to implement the integer overflow checks in a much more
3195        dynamic way, which I have now done. The artificial limitation on group
3196        length has been removed - we now have only the limit on the total length of
3197        the compiled pattern, which depends on the LINK_SIZE setting.
3198    
3199    10. Fixed a bug in the documentation for get/copy named substring when
3200        duplicate names are permitted. If none of the named substrings are set, the
3201        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3202        empty string.
3203    
3204    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3205        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3206        because the ] is interpreted as the first data character and the
3207        terminating ] is not found. PCRE has been made compatible with Perl in this
3208        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3209        cause memory overwriting.
3210    
3211    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3212        string has been matched (to stop an infinite loop). It was not recognizing
3213        a conditional subpattern that could match an empty string if that
3214        subpattern was within another subpattern. For example, it looped when
3215        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3216        condition was not nested. This bug has been fixed.
3217    
3218    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3219        past the start of the subject in the presence of bytes with the top bit
3220        set, for example "\x8aBCD".
3221    
3222    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3223        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3224    
3225    14. Optimized (?!) to (*FAIL).
3226    
3227    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3228        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3229        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3230        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3231        does: it's just the validity check that is more restrictive.
3232    
3233    16. Inserted checks for integer overflows during escape sequence (backslash)
3234        processing, and also fixed erroneous offset values for syntax errors during
3235        backslash processing.
3236    
3237    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3238        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3239    
3240    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3241        caused an overrun.
3242    
3243    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3244        something other than just ASCII characters) inside a group that had an
3245        unlimited repeat caused a loop at compile time (while checking to see
3246        whether the group could match an empty string).
3247    
3248    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3249        [\P{Any}] did so. (Error in the code for printing property names.)
3250    
3251    21. An orphan \E inside a character class could cause a crash.
3252    
3253    22. A repeated capturing bracket such as (A)? could cause a wild memory
3254        reference during compilation.
3255    
3256    23. There are several functions in pcre_compile() that scan along a compiled
3257        expression for various reasons (e.g. to see if it's fixed length for look
3258        behind). There were bugs in these functions when a repeated \p or \P was
3259        present in the pattern. These operators have additional parameters compared
3260        with \d, etc, and these were not being taken into account when moving along
3261        the compiled data. Specifically:
3262    
3263        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3264            length.
3265    
3266        (b) An item such as \pL+ within a repeated group could cause crashes or
3267            loops.
3268    
3269        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3270            "reference to non-existent subpattern" error.
3271    
3272        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3273    
3274    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3275        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3276    
3277    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3278    
3279    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3280        character were causing crashes (broken optimization).
3281    
3282    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3283        \p or \P) caused a compile-time loop.
3284    
3285    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3286        break. For example, the unstudied pattern [\r\n]A does not match the string
3287        "\r\nA" because change 7.0/46 below moves the current point on by two
3288        characters after failing to match at the start. However, the pattern \nA
3289        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3290        the same is true. There doesn't seem any very clean way out of this, but
3291        what I have chosen to do makes the common cases work: PCRE now takes note
3292        of whether there can be an explicit match for \r or \n anywhere in the
3293        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3294        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3295        pattern has explicit CR or LF references.
3296    
3297    29. Added (*CR) etc for changing newline setting at start of pattern.
3298    
3299    
3300    Version 7.2 19-Jun-07
3301  ---------------------  ---------------------
3302    
3303   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3317  Version 7.2 01-May-07
3317      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3318      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3319      in all cases.      in all cases.
3320    
3321   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3322    
3323      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3324          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3325    
3326      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3327          to be opened parentheses.          to be opened parentheses.
3328    
3329      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3330          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3331    
3332        (d) \K resets the start of the current match so that everything before
3333            is not part of it.
3334    
3335        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3336    
3337        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3338            reference syntax.
3339    
3340        (g) (?| introduces a group in which the numbering of parentheses in each
3341            alternative starts with the same number.
3342    
3343        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3344    
3345     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3346        PCRE_INFO_JCHANGED.
3347    
3348     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3349        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3350        for detecting groups that can match an empty string.
3351    
3352     9. A pattern with a very large number of alternatives (more than several
3353        hundred) was running out of internal workspace during the pre-compile
3354        phase, where pcre_compile() figures out how much memory will be needed. A
3355        bit of new cunning has reduced the workspace needed for groups with
3356        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3357        workspace instead of running out of the 4096 that are available.
3358    
3359    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3360    
3361    11. Applied patch from Google to remove an optimization that didn't quite work.
3362        The report of the bug said:
3363    
3364          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3365          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3366          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3367    
3368    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3369        it matched the wrong number of bytes.
3370    
3371    
3372  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3373  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1680